{"series":"The Gateway Charts","canonical":"https://www.requesty.ai/data/weekly/2026-w46","license":"CC BY 4.0","citation":"Source: Requesty, The Gateway Charts, week 46 (Nov 9 to 15, 2026). https://www.requesty.ai/data/weekly/2026-w46","id":"2026-w46","year":2026,"week":46,"start":"2026-11-09","end":"2026-11-15","compare":{"start":"2026-11-02","end":"2026-11-08"},"weeks":["2026-08-24","2026-08-31","2026-09-07","2026-09-14","2026-09-21","2026-09-28","2026-10-05","2026-10-12","2026-10-19","2026-10-26","2026-11-02","2026-11-09"],"published":null,"status":"draft","title":"Last week in AI. Who climbed. Who slipped.","dek":"Every request through the Requesty gateway names a model, a host and the app that sent it. Each week we rank them against the week before and pull out what changed, and what it means for what you build.","findings":[{"id":"top-model","topic":"Rankings","claim":"deepseek-v4.1-flash held #1 on the Requesty gateway in week 46 (Nov 9 to 15, 2026) with 28.3% of tokens, up 2.27 points on the week before.","short":"deepseek-v4.1-flash pulled further ahead.","soWhat":"Last week it gave up share for the first time since September. This week it took it back, and it now carries more than the next three models combined. The order below it is unchanged: claude-opus-5-5 at 10.5%, gpt-6-luna at 8.5%.","action":{"text":"See your own model mix by key and team","href":"https://docs.requesty.ai/features/usage-analytics"},"metric":{"value":28.3,"unit":"pct","denominator":"of tokens, each team capped at 2%","prev":26.03,"better":"up"},"footnote":"Share of tokens by model · Nov 9 to 15, 2026 vs Nov 2 to 8","figure":{"kind":"ranking","ranking":"models","top":10,"highlight":"deepseek-v4-flash"},"posts":{"x":["deepseek-v4.1-flash was the most used model on Requesty again last week (Nov 9 to 15): 28.3% of tokens, up 2.27 pts. That's more than the next three models combined.","Top 3 models on Requesty last week, by share of tokens: deepseek-v4.1-flash 28.3%, claude-opus-5-5 10.5%, gpt-6-luna 8.5%. Same order for the third week running.","One model carries more than a quarter of the tokens on Requesty: deepseek-v4.1-flash, 28.3% last week (Nov 9 to 15). #1 in 8 of the last 12 weeks."],"linkedin":"deepseek-v4.1-flash held #1 on the Requesty gateway in week 46 (Nov 9 to 15, 2026) with 28.3% of tokens, up 2.27 points on the week before.\n\nLast week it gave up share for the first time since September. This week it took it back, and it now carries more than the next three models combined. The order below it is unchanged: claude-opus-5-5 at 10.5%, gpt-6-luna at 8.5%.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"url":"https://www.requesty.ai/data/weekly/2026-w46/top-model","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/top-model/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/top-model/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/top-model/landscape.png"}},{"id":"coding-agents","topic":"Apps","claim":"Five coding agents (OpenCode, Claude Code, Pi, DeepSeek Harness and GitHub Copilot) sent 33.4% of tokens through the Requesty gateway in week 46 (Nov 9 to 15, 2026).","short":"Coding agents sent a third of tokens.","soWhat":"That is the second week in a row at about a third. The SDKs in the top 10 send 38.3% between them, so agents are now almost level with every SDK and framework we can name. Their requests are long, cached and end in tool calls.","action":{"text":"Run Claude Code through Requesty","href":"https://docs.requesty.ai/integrations/claude-code"},"metric":{"value":33.4,"unit":"pct","denominator":"of tokens, each team capped at 2%"},"footnote":"Share of tokens by calling app, team-capped · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"pct","rows":[{"label":"OpenAI Python SDK","value":21.27,"prev":19.11,"highlight":false,"sub":"SDK or framework"},{"label":"OpenCode","value":13.14,"prev":14.14,"highlight":true,"sub":"Coding agent"},{"label":"Vercel AI SDK","value":10.06,"prev":9.29,"highlight":false,"sub":"SDK or framework"},{"label":"Claude Code","value":9.83,"prev":11.48,"highlight":true,"sub":"Coding agent"},{"label":"Pi","value":6.64,"prev":3.53,"highlight":true,"sub":"Coding agent"},{"label":"Anthropic SDK","value":2.51,"prev":2.67,"highlight":false,"sub":"SDK or framework"},{"label":"OpenAI SDK (other)","value":2.36,"prev":2.49,"highlight":false,"sub":"SDK or framework"},{"label":"DeepSeek Harness","value":2.15,"prev":2.69,"highlight":true,"sub":"Coding agent"},{"label":"LiteLLM","value":2.11,"prev":1.9,"highlight":false,"sub":"SDK or framework"},{"label":"GitHub Copilot","value":1.6,"prev":1.93,"highlight":true,"sub":"Coding agent"}]},"posts":{"x":["About a third of all tokens on Requesty last week came from five coding agents: OpenCode 13.1%, Claude Code 9.8%, Pi 6.6%, DeepSeek Harness 2.2% and GitHub Copilot 1.6%. Together 33.4% (Nov 9 to 15).","OpenAI's Python SDK still sends the most tokens through Requesty (21.3% last week). The five coding agents in the top 10 add up to 33.4%, close to the SDKs in the top 10 combined (38.3%).","Pi is the coding agent to watch on Requesty: 6.6% of tokens last week (Nov 9 to 15), up from 3.5% the week before. Five agents now send 33.4% of all tokens."],"linkedin":"Five coding agents (OpenCode, Claude Code, Pi, DeepSeek Harness and GitHub Copilot) sent 33.4% of tokens through the Requesty gateway in week 46 (Nov 9 to 15, 2026).\n\nThat is the second week in a row at about a third. The SDKs in the top 10 send 38.3% between them, so agents are now almost level with every SDK and framework we can name. Their requests are long, cached and end in tool calls.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"url":"https://www.requesty.ai/data/weekly/2026-w46/coding-agents","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/coding-agents/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/coding-agents/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/coding-agents/landscape.png"}},{"id":"multi-model","topic":"Usage","claim":"67.6% of active teams on the Requesty gateway used two or more models in week 46 (Nov 9 to 15, 2026); 58.6% used two or more labs and 60.2% two or more providers.","short":"Most teams used more than one model.","soWhat":"Single-model stacks are still the minority. Almost one team in five ran six models or more, so switching and mixing models is routine work, not a migration project.","action":{"text":"Switch models without changing your integration","href":"https://docs.requesty.ai/quickstart"},"metric":{"value":67.6,"unit":"pct","denominator":"of active teams","prev":69},"footnote":"Share of active teams · Nov 9 to 15, 2026","figure":{"kind":"stacked","bins":["1","2","3","4 to 5","6 or more"],"rows":[{"label":"Models","values":[32.4,19,13.6,16.5,18.6],"headline":67.6,"headlineLabel":"use 2 or more"},{"label":"Labs","values":[41.4,25.4,13.7,12.2,7.3],"headline":58.6,"headlineLabel":"use 2 or more"},{"label":"Providers","values":[39.8,26.3,12.2,12.3,9.4],"headline":60.2,"headlineLabel":"use 2 or more"}]},"posts":{"x":["67.6% of active teams on Requesty used two or more models last week (Nov 9 to 15). 58.6% used two or more labs, and 60.2% two or more providers. Almost one in five ran six models or more.","Only 32% of teams on Requesty used a single model last week. The rest mix: 19% used two, 14% three, and 35% four or more. Plan for more than one model from day one.","6 in 10 teams on Requesty sent traffic to more than one provider last week (60.2%, fallbacks and routing policies count too). Switching models and hosts is normal now."],"linkedin":"67.6% of active teams on the Requesty gateway used two or more models in week 46 (Nov 9 to 15, 2026); 58.6% used two or more labs and 60.2% two or more providers.\n\nSingle-model stacks are still the minority. Almost one team in five ran six models or more, so switching and mixing models is routine work, not a migration project.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"url":"https://www.requesty.ai/data/weekly/2026-w46/multi-model","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/multi-model/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/multi-model/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/multi-model/landscape.png"}},{"id":"speed-up","topic":"Speed","claim":"claude-haiku-5-5 reached its first token 1.3x sooner on the Requesty gateway in week 46 (Nov 9 to 15, 2026): a median of 1.76 s, down from 2.21 s the week before, on streaming requests.","short":"claude-haiku-5-5 started 1.3× sooner.","soWhat":"It started sooner but streamed slower: its median output speed fell from 462 to 291 tokens per second. First token and stream speed can move in opposite directions, so measure the one your users feel.","action":{"text":"Let the gateway pick the fastest host","href":"https://docs.requesty.ai/features/latency-routing"},"metric":{"value":1.3,"unit":"x","denominator":"median time to first token, this week vs last","better":"up"},"footnote":"Median first token, streaming · Nov 9 to 15, 2026 vs Nov 2 to 8","figure":{"kind":"speedTape","rows":[{"name":"claude-haiku-5-5","sub":"Anthropic","from":2206,"to":1755,"highlight":true},{"name":"deepseek-v4-flash","sub":"DeepSeek","from":1432,"to":1140,"highlight":false},{"name":"deepseek-v4.1-flash","sub":"DeepSeek","from":972,"to":834,"highlight":false},{"name":"glm-5.3","sub":"Z.ai","from":842,"to":737,"highlight":false},{"name":"leanstral-1-5","sub":"Mistral","from":1847,"to":1659,"highlight":false}]},"posts":{"x":["Speed-up of the week on Requesty: claude-haiku-5-5. Median time to first token went from 2.21 s to 1.76 s (Nov 9 to 15 vs the week before), 1.3x sooner. Measured at the gateway on streaming requests.","claude-haiku-5-5 sent its first token 1.3x sooner on Requesty last week, but its median stream speed fell from 462 to 291 tok/s. Faster to start isn't always faster to finish.","The biggest first-token speed-up in our top 20 models last week (Nov 9 to 15): claude-haiku-5-5, 2.21 s to 1.76 s at the median. The biggest slowdown: claude-sonnet-5, 1.65 s to 2.32 s."],"linkedin":"claude-haiku-5-5 reached its first token 1.3x sooner on the Requesty gateway in week 46 (Nov 9 to 15, 2026): a median of 1.76 s, down from 2.21 s the week before, on streaming requests.\n\nIt started sooner but streamed slower: its median output speed fell from 462 to 291 tokens per second. First token and stream speed can move in opposite directions, so measure the one your users feel.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"caveats":["Latency is measured at the gateway, so it includes the host and the network, and reflects the prompts teams sent that week.","Measured at the gateway on streaming requests, so it blends the model, the host and the prompts teams sent that week."],"url":"https://www.requesty.ai/data/weekly/2026-w46/speed-up","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/speed-up/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/speed-up/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/speed-up/landscape.png"}},{"id":"rescues","topic":"Reliability","claim":"In week 46 (Nov 9 to 15, 2026), 96.1% of requests on a fallback policy whose first attempt hit a provider error were still served by another host on the Requesty gateway, up from 86.4% the week before.","short":"96% of failed requests still got served.","soWhat":"A provider error only reaches your users when there is nowhere else to send the request. This week most rescues ran from Bedrock to Anthropic's own API, the same model on another host.","action":{"text":"Set up a fallback policy","href":"https://docs.requesty.ai/features/fallback-policies"},"metric":{"value":96.1,"unit":"pct","denominator":"of requests on a fallback policy whose first host returned a provider error","prev":86.4,"better":"up"},"footnote":"Requests whose first host failed · Nov 9 to 15, 2026","figure":{"kind":"split","total":200,"parts":[{"label":"Still served","pct":96.1,"state":"ok"},{"label":"Still failed","pct":3.9,"state":"failed"}],"stats":[{"label":"Fallback policy","value":96.1,"unit":"pct"},{"label":"Latency policy","value":79,"unit":"pct"},{"label":"Same model served","value":79.1,"unit":"pct"}]},"posts":{"x":["When a request's first host failed on Requesty last week, a fallback policy still got it served 96.1% of the time (Nov 9 to 15). The week before: 86.4%.","A provider error doesn't have to reach your app. Last week 96.1% of requests whose first host failed were still served through a fallback policy on Requesty (Nov 9 to 15).","Busiest rescue lane on Requesty last week: Bedrock to Anthropic, 63.6% of all rescues. Same model, different host, and the user never sees the error (Nov 9 to 15)."],"linkedin":"In week 46 (Nov 9 to 15, 2026), 96.1% of requests on a fallback policy whose first attempt hit a provider error were still served by another host on the Requesty gateway, up from 86.4% the week before.\n\nA provider error only reaches your users when there is nowhere else to send the request. This week most rescues ran from Bedrock to Anthropic's own API, the same model on another host.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"url":"https://www.requesty.ai/data/weekly/2026-w46/rescues","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/rescues/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/rescues/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/rescues/landscape.png"}},{"id":"prompt-cache","topic":"Caching","claim":"78.1% of input tokens through the Requesty gateway in week 46 (Nov 9 to 15, 2026) were read from a prompt cache; models read 46 input tokens for every output token.","short":"78.1% of input came from cache.","soWhat":"Input is most of the bill for agent workloads, and the cache decides what that input costs. Compare models on what you pay after caching, not on the list price for output.","action":{"text":"Turn on auto caching","href":"https://docs.requesty.ai/features/auto-caching"},"metric":{"value":78.1,"unit":"pct","denominator":"of input tokens","prev":77.1,"better":"up"},"footnote":"Share of input tokens read from a prompt cache · Nov 9 to 15, 2026","figure":{"kind":"trend","unit":"pct","series":[{"label":"Input read from cache","values":[81.7,76.3,71.9,74.9,73.6,76.3,75.3,74.4,77.1,78.1,77.1,78.1],"highlight":true}]},"posts":{"x":["78.1% of input tokens on Requesty last week were read from a prompt cache (Nov 9 to 15). Models read 46 input tokens for every token they wrote. Output price is the smaller part of most bills.","For every token a model wrote on Requesty last week, it read 46. 78.1% of that input came from a prompt cache (Nov 9 to 15). Check your cache hit rate before you compare price lists.","Prompt caching on Requesty, 12 weeks: as low as 71.9% of input tokens in early September, 78.1% last week (Nov 9 to 15). The workload is context, replayed on every turn."],"linkedin":"78.1% of input tokens through the Requesty gateway in week 46 (Nov 9 to 15, 2026) were read from a prompt cache; models read 46 input tokens for every output token.\n\nInput is most of the bill for agent workloads, and the cache decides what that input costs. Compare models on what you pay after caching, not on the list price for output.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."},"url":"https://www.requesty.ai/data/weekly/2026-w46/prompt-cache","cards":{"square":"https://www.requesty.ai/data/weekly/2026-w46/prompt-cache/square.png","portrait":"https://www.requesty.ai/data/weekly/2026-w46/prompt-cache/portrait.png","landscape":"https://www.requesty.ai/data/weekly/2026-w46/prompt-cache/landscape.png"}}],"rankings":[{"id":"models","label":"Models","unit":"pct","denominator":"of tokens, each team capped at 2%","rows":[{"rank":1,"last":1,"name":"deepseek-v4.1-flash","sub":"DeepSeek","share":28.3,"prevShare":26.03,"peak":1,"weeks":10,"series":[0,0,6.44,19.08,20.37,14.63,17.8,20.41,26.23,29.56,26.03,28.3],"ranks":[null,null,4,1,1,2,1,1,1,1,1,1]},{"rank":2,"last":2,"name":"claude-opus-5-5","sub":"Anthropic","share":10.52,"prevShare":10.24,"peak":1,"weeks":8,"series":[0,0,0,0,1.68,14.82,10.85,7.06,8.61,7.78,10.24,10.52],"ranks":[null,null,null,null,13,1,2,4,2,2,2,2]},{"rank":3,"last":3,"name":"gpt-6-luna","sub":"OpenAI","share":8.46,"prevShare":8.33,"peak":3,"weeks":8,"series":[0,0,0,0,4.92,9.93,9.87,8.14,7.4,7.4,8.33,8.46],"ranks":[null,null,null,null,6,4,3,3,4,3,3,3]},{"rank":4,"last":4,"name":"glm-5.3-flash","sub":"Z.ai","share":5.42,"prevShare":6.04,"peak":2,"weeks":12,"series":[3.37,9.61,11.52,11.94,10.03,10.88,7.77,7,5.69,4.28,6.04,5.42],"ranks":[10,2,2,3,3,3,5,5,5,5,4,4]},{"rank":5,"last":6,"name":"claude-sonnet-5-5","sub":"Anthropic","share":5.35,"prevShare":5.27,"peak":5,"weeks":7,"series":[0,0,0,0,0,1.29,3.79,2.99,3.46,3.85,5.27,5.35],"ranks":[null,null,null,null,null,13,8,8,8,7,6,5]},{"rank":6,"last":5,"name":"gpt-6.1-sol","sub":"OpenAI","share":5.04,"prevShare":5.4,"peak":4,"weeks":7,"series":[0,0,0,0,0,1.46,4.09,4.24,5.41,5.95,5.4,5.04],"ranks":[null,null,null,null,null,11,6,7,6,4,5,6]},{"rank":7,"last":7,"name":"deepseek-v4-flash","sub":"DeepSeek","share":4.57,"prevShare":4.26,"peak":1,"weeks":12,"series":[19.2,19.05,17.55,13.53,11.81,8.68,8.78,9.2,8.26,4.22,4.26,4.57],"ranks":[1,1,1,2,2,5,4,2,3,6,7,7]},{"rank":8,"last":8,"name":"claude-haiku-5-5","sub":"Anthropic","share":3.75,"prevShare":3.5,"peak":8,"weeks":6,"series":[0,0,0,0,0,0,0.87,1.69,2.48,3.55,3.5,3.75],"ranks":[null,null,null,null,null,null,18,12,9,8,8,8]},{"rank":9,"last":10,"name":"nemotron-3-ultra-550b-a55b","sub":"NVIDIA","share":2.4,"prevShare":2.37,"peak":9,"weeks":12,"series":[2.25,2.04,1.1,1.55,1.85,1.56,1.25,1.03,2.21,2.78,2.37,2.4],"ranks":[12,11,19,11,10,10,13,14,11,10,10,9]},{"rank":10,"last":11,"name":"claude-sonnet-5","sub":"Anthropic","share":2.18,"prevShare":2.01,"peak":4,"weeks":12,"series":[6.37,7.6,5.07,3.89,6.22,4.56,3.94,2.75,2.42,3.21,2.01,2.18],"ranks":[4,4,5,6,4,6,7,10,10,9,11,10]},{"rank":11,"last":13,"name":"gpt-6.1-terra","sub":"OpenAI","share":2.16,"prevShare":1.34,"peak":11,"weeks":3,"series":[0,0,0,0,0,0,0,0,0,1.37,1.34,2.16],"ranks":[null,null,null,null,null,null,null,null,null,14,13,11]},{"rank":12,"last":9,"name":"gpt-5.6-luna","sub":"OpenAI","share":2.02,"prevShare":2.39,"peak":3,"weeks":12,"series":[8.46,9.46,8.88,9.59,5.08,3.71,2.86,2.76,2.05,2.33,2.39,2.02],"ranks":[3,3,3,4,5,7,9,9,12,11,9,12]},{"rank":13,"last":14,"name":"gemini-3.8-flash","sub":"Google","share":1.45,"prevShare":1.29,"peak":6,"weeks":11,"series":[0,0.65,1.13,1.36,1.05,1,2.49,6.7,4.4,0.92,1.29,1.45],"ranks":[null,20,18,14,20,17,10,6,7,16,14,13]},{"rank":14,"last":12,"name":"nemotron-3.5-lightning-30b-a3b","sub":"NVIDIA","share":1.11,"prevShare":1.52,"peak":12,"weeks":12,"series":[0.16,0.17,0.22,0.26,0.14,0.26,0.39,0.42,1.2,1.37,1.52,1.11],"ranks":[32,34,32,28,42,31,26,27,14,13,12,14]},{"rank":15,"last":15,"name":"glm-5.3","sub":"Z.ai","share":0.88,"prevShare":1.23,"peak":7,"weeks":12,"series":[1.22,2.59,2.77,3,3.19,3.27,2.27,2.01,1.18,1.22,1.23,0.88],"ranks":[17,10,8,7,7,8,11,11,15,15,15,15]},{"rank":16,"last":null,"name":"gemini-4-carbon","sub":"Google","share":0.73,"prevShare":0,"peak":16,"weeks":1,"series":[0,0,0,0,0,0,0,0,0,0,0,0.73],"ranks":[null,null,null,null,null,null,null,null,null,null,null,16]},{"rank":17,"last":17,"name":"claude-opus-5","sub":"Anthropic","share":0.59,"prevShare":0.73,"peak":5,"weeks":12,"series":[5.1,3.64,3.78,6.03,2.42,1.11,0.8,0.49,0.54,0.42,0.73,0.59],"ranks":[7,7,6,5,9,16,19,22,21,25,17,17]},{"rank":18,"last":19,"name":"claude-sonnet-4-6","sub":"Anthropic","share":0.58,"prevShare":0.57,"peak":13,"weeks":12,"series":[2.16,1.75,1.44,1.28,1.08,0.85,1.02,0.87,0.67,0.5,0.57,0.58],"ranks":[13,15,16,15,19,18,15,15,17,23,19,18]},{"rank":19,"last":23,"name":"minimax-m3.1-flash","sub":"MiniMax","share":0.56,"prevShare":0.51,"peak":12,"weeks":3,"series":[0,0,0,0,0,0,0,0,0.68,1.62,0.51,0.56],"ranks":[null,null,null,null,null,null,null,null,null,12,23,19]},{"rank":20,"last":24,"name":"leanstral-1-5","sub":"Mistral","share":0.53,"prevShare":0.5,"peak":9,"weeks":12,"series":[1.44,1.94,2.54,2.09,0.7,0.43,0.3,0.46,0.59,0.7,0.5,0.53],"ranks":[16,13,9,9,21,24,29,25,18,20,24,20]},{"rank":21,"last":16,"name":"ling-3.1-flash","sub":"Other","share":0.52,"prevShare":0.92,"peak":16,"weeks":6,"series":[0,0,0,0,0,0,0.09,0.22,0.31,0.82,0.92,0.52],"ranks":[null,null,null,null,null,null,44,33,27,18,16,21]},{"rank":22,"last":22,"name":"gemini-4-argon","sub":"Google","share":0.47,"prevShare":0.51,"peak":22,"weeks":5,"series":[0,0,0,0,0,0,0,0.47,0.43,0.45,0.51,0.47],"ranks":[null,null,null,null,null,null,null,24,26,24,22,22]},{"rank":23,"last":18,"name":"deepseek-v4-pro","sub":"DeepSeek","share":0.43,"prevShare":0.72,"peak":5,"weeks":12,"series":[6.31,2.62,1.51,0.75,1.24,1.15,1.03,0.78,0.83,0.9,0.72,0.43],"ranks":[5,9,15,18,17,14,14,16,16,17,18,23]},{"rank":24,"last":21,"name":"gpt-6-sol","sub":"OpenAI","share":0.38,"prevShare":0.55,"peak":9,"weeks":8,"series":[0,0,0,0,1.72,1.58,0.92,0.56,0.57,0.53,0.55,0.38],"ranks":[null,null,null,null,11,9,17,20,19,22,21,24]},{"rank":25,"last":20,"name":"gemini-3.5-flash-lite","sub":"Google","share":0.31,"prevShare":0.56,"peak":18,"weeks":11,"series":[0.66,0.59,0.62,0.66,0.5,0.21,0.55,0.72,0.55,0.29,0.56,0.31],"ranks":[23,22,null,19,24,36,24,18,20,29,20,25]},{"rank":26,"last":26,"name":"gemini-3.1-flash-lite","sub":"Google","share":0.24,"prevShare":0.38,"peak":17,"weeks":12,"series":[0.48,0.55,0.54,0.75,0.48,0.35,0.37,0.43,0.29,0.24,0.38,0.24],"ranks":[24,25,24,17,25,26,27,26,28,30,26,26]},{"rank":27,"last":28,"name":"gemini-2.5-flash-lite","sub":"Google","share":0.24,"prevShare":0.25,"peak":25,"weeks":12,"series":[0.36,0.44,0.52,0.34,0.23,0.29,0.14,0.27,0.23,0.42,0.25,0.24],"ranks":[28,27,25,25,32,30,40,32,32,26,28,27]},{"rank":28,"last":25,"name":"kimi-k3.1-pro","sub":"Moonshot","share":0.22,"prevShare":0.44,"peak":25,"weeks":4,"series":[0,0,0,0,0,0,0,0,0.03,0.3,0.44,0.22],"ranks":[null,null,null,null,null,null,null,null,55,28,25,28]},{"rank":29,"last":null,"name":"gemini-4-hydrogen","sub":"Google","share":0.2,"prevShare":0,"peak":29,"weeks":1,"series":[0,0,0,0,0,0,0,0,0,0,0,0.2],"ranks":[null,null,null,null,null,null,null,null,null,null,null,29]},{"rank":30,"last":29,"name":"gemini-3.7-flash","sub":"Google","share":0.19,"prevShare":0.25,"peak":8,"weeks":12,"series":[4.29,1.1,0.43,0.35,0.22,0.33,0.2,0.08,0.12,0.11,0.25,0.19],"ranks":[8,17,27,24,35,28,37,45,40,41,29,30]},{"rank":31,"last":32,"name":"claude-haiku-4-5","sub":"Anthropic","share":0.14,"prevShare":0.21,"peak":19,"weeks":12,"series":[0.77,0.9,0.76,0.64,0.67,0.77,0.74,0.35,0.22,0.23,0.21,0.14],"ranks":[20,19,22,20,22,20,20,29,33,32,32,31]},{"rank":32,"last":35,"name":"gemma-4-31b-it","sub":"Google","share":0.14,"prevShare":0.17,"peak":32,"weeks":9,"series":[0.27,0.12,0.19,0.18,0.16,0.13,0.14,0.11,0.45,0.49,0.17,0.14],"ranks":[null,39,36,34,39,41,39,42,null,null,35,32]},{"rank":33,"last":34,"name":"gemini-3.1-pro-preview","sub":"Google","share":0.11,"prevShare":0.19,"peak":33,"weeks":3,"series":[0.23,0.3,0.33,0.13,0.07,0.06,0.06,0.17,0.22,0.14,0.19,0.11],"ranks":[null,null,null,null,null,null,47,null,null,null,34,33]},{"rank":34,"last":37,"name":"gemini-2.5-flash","sub":"Google","share":0.1,"prevShare":0.14,"peak":32,"weeks":7,"series":[1.56,1.9,2.48,2.37,2.49,0.23,0.26,0.21,0.19,0.12,0.14,0.1],"ranks":[null,null,null,null,null,35,32,35,34,38,37,34]},{"rank":35,"last":38,"name":"glm-5.2","sub":"Z.ai","share":0.1,"prevShare":0.11,"peak":6,"weeks":5,"series":[5.18,3.6,3.04,1.88,1.28,0.67,0.41,0.47,0.15,0.21,0.11,0.1],"ranks":[6,8,null,null,null,null,null,23,null,null,38,35]},{"rank":36,"last":36,"name":"gpt-5.4","sub":"OpenAI","share":0.09,"prevShare":0.16,"peak":18,"weeks":12,"series":[0.75,0.92,0.51,0.36,0.3,0.24,0.3,0.21,0.15,0.17,0.16,0.09],"ranks":[21,18,26,23,28,33,30,34,38,35,36,36]},{"rank":37,"last":31,"name":"gpt-5.6-sol","sub":"OpenAI","share":0.09,"prevShare":0.22,"peak":2,"weeks":12,"series":[8.55,5.49,2.43,2.27,1.4,0.83,0.31,0.27,0.17,0.09,0.22,0.09],"ranks":[2,5,10,8,16,19,28,31,35,42,31,37]},{"rank":38,"last":33,"name":"mistral-large-4","sub":"Mistral","share":0.09,"prevShare":0.2,"peak":31,"weeks":6,"series":[0,0,0,0,0,0,0.27,0.07,0.07,0.2,0.2,0.09],"ranks":[null,null,null,null,null,null,31,46,47,34,33,38]},{"rank":39,"last":40,"name":"gpt-5.5","sub":"OpenAI","share":0.09,"prevShare":0.09,"peak":28,"weeks":11,"series":[0.27,0.2,0.34,0.22,0.09,0.11,0.05,0.31,0.25,0.07,0.09,0.09],"ranks":[30,32,28,31,47,46,51,null,30,45,40,39]},{"rank":40,"last":39,"name":"gemini-3.5-flash","sub":"Google","share":0.08,"prevShare":0.1,"peak":27,"weeks":12,"series":[0.32,0.43,0.31,0.29,0.2,0.15,0.24,0.14,0.1,0.14,0.1,0.08],"ranks":[29,28,29,27,38,38,34,41,42,36,39,40]}]},{"id":"apps","label":"Apps","unit":"pct","denominator":"of tokens, each team capped at 2%","rows":[{"rank":1,"last":1,"name":"OpenAI Python SDK","sub":"SDK or framework","share":21.27,"prevShare":19.11,"peak":1,"weeks":12,"series":[22.5,23.4,20.15,22.14,21.04,20.08,19.88,18.14,18.48,18.9,19.11,21.27],"ranks":[1,1,1,1,1,1,1,1,1,1,1,1]},{"rank":2,"last":2,"name":"OpenCode","sub":"Coding agent","share":13.14,"prevShare":14.14,"peak":2,"weeks":12,"series":[12.64,13.85,14.19,12.66,16.78,12.65,10.89,12.02,12.81,15.42,14.14,13.14],"ranks":[2,2,2,2,2,2,2,2,2,2,2,2]},{"rank":3,"last":4,"name":"Vercel AI SDK","sub":"SDK or framework","share":10.06,"prevShare":9.29,"peak":3,"weeks":12,"series":[6.73,7.06,8.93,8.82,9.08,7.98,7.54,6.5,8.8,8.32,9.29,10.06],"ranks":[4,4,3,3,3,3,4,4,4,4,4,3]},{"rank":4,"last":3,"name":"Claude Code","sub":"Coding agent","share":9.83,"prevShare":11.48,"peak":3,"weeks":12,"series":[6.14,5.73,8.36,4.24,5.6,6.07,9.24,11.73,11.37,10.54,11.48,9.83],"ranks":[5,5,4,6,6,4,3,3,3,3,3,4]},{"rank":5,"last":5,"name":"Pi","sub":"Coding agent","share":6.64,"prevShare":3.53,"peak":4,"weeks":12,"series":[2.88,4.71,6.8,7.65,6.27,4.33,5.39,4.1,4.54,4.01,3.53,6.64],"ranks":[7,6,5,4,5,6,5,6,5,5,5,5]},{"rank":6,"last":7,"name":"Anthropic SDK","sub":"SDK or framework","share":2.51,"prevShare":2.67,"peak":6,"weeks":9,"series":[1.24,0.33,1.37,0.28,0.63,0.69,1.06,0.64,0.65,1.16,2.67,2.51],"ranks":[9,12,null,14,null,10,null,12,11,10,7,6]},{"rank":7,"last":8,"name":"OpenAI SDK (other)","sub":"SDK or framework","share":2.36,"prevShare":2.49,"peak":3,"weeks":12,"series":[9.13,7.64,4.62,5.39,6.48,3.11,2.82,2.82,2.68,2.92,2.49,2.36],"ranks":[3,3,6,5,4,7,7,7,6,6,8,7]},{"rank":8,"last":6,"name":"DeepSeek Harness","sub":"Coding agent","share":2.15,"prevShare":2.69,"peak":6,"weeks":11,"series":[0.82,0.83,1.25,1.94,1.28,0.63,0.76,0.97,1.37,2.01,2.69,2.15],"ranks":[10,9,10,9,null,12,12,9,8,8,6,8]},{"rank":9,"last":10,"name":"LiteLLM","sub":"SDK or framework","share":2.11,"prevShare":1.9,"peak":9,"weeks":11,"series":[0.79,0.43,0.53,0.95,0.71,0.74,1.53,0.95,1.23,1.39,1.9,2.11],"ranks":[null,11,12,11,12,9,9,10,9,9,10,9]},{"rank":10,"last":9,"name":"GitHub Copilot","sub":"Coding agent","share":1.6,"prevShare":1.93,"peak":6,"weeks":12,"series":[3.06,2.96,4.01,2.83,1.91,1.96,1.67,1.8,1.84,2.7,1.93,1.6],"ranks":[6,7,7,7,8,8,8,8,7,7,9,10]},{"rank":11,"last":11,"name":"Kilo Code","sub":"Coding agent","share":1.39,"prevShare":1.27,"peak":8,"weeks":12,"series":[1.74,1.35,1.91,1.09,0.99,0.61,1.13,0.89,1.11,0.83,1.27,1.39],"ranks":[8,8,8,10,9,13,11,11,10,11,11,11]},{"rank":12,"last":14,"name":"Zoo Code","sub":"Coding agent","share":0.27,"prevShare":0.18,"peak":11,"weeks":12,"series":[0.48,0.18,0.33,0.66,0.73,0.63,0.57,0.42,0.34,0.18,0.18,0.27],"ranks":[11,16,14,12,11,11,13,14,13,15,14,12]},{"rank":13,"last":13,"name":"Open WebUI","sub":"App","share":0.25,"prevShare":0.26,"peak":13,"weeks":12,"series":[0.19,0.25,0.17,0.15,0.29,0.2,0.2,0.21,0.2,0.21,0.26,0.25],"ranks":[14,14,17,16,14,15,15,16,15,14,13,13]},{"rank":14,"last":12,"name":"LangChain","sub":"SDK or framework","share":0.21,"prevShare":0.3,"peak":10,"weeks":12,"series":[0.42,0.48,0.42,0.34,0.28,0.28,0.31,0.25,0.24,0.26,0.3,0.21],"ranks":[12,10,13,13,15,14,14,15,14,13,12,14]},{"rank":15,"last":17,"name":"LoreBary","sub":"App","share":0.01,"prevShare":0.01,"peak":15,"weeks":10,"series":[0.01,0.01,0.01,0.01,0.01,0.01,0.01,0.01,0.01,0.01,0.01,0.01],"ranks":[null,17,18,18,17,18,null,18,16,16,17,15]}]},{"id":"labs","label":"Labs","unit":"pct","denominator":"of tokens, each team capped at 2%","rows":[{"rank":1,"last":1,"name":"DeepSeek","share":34.07,"prevShare":32.03,"peak":1,"weeks":12,"series":[26.65,23,26.91,34.37,34.37,25.39,28.5,31.16,36.15,35.51,32.03,34.07],"ranks":[1,1,1,1,1,2,1,1,1,1,1,1]},{"rank":2,"last":2,"name":"Anthropic","share":24.4,"prevShare":23.9,"peak":1,"weeks":12,"series":[19.29,18.76,17.34,14.9,14.76,25.65,24.22,18.3,20.26,20.84,23.9,24.4],"ranks":[3,3,2,4,3,1,2,3,2,2,2,2]},{"rank":3,"last":3,"name":"OpenAI","share":19.69,"prevShare":20.2,"peak":2,"weeks":12,"series":[21.89,20.14,16.91,16.29,18.34,20.72,20.79,18.57,18.31,19.68,20.2,19.69],"ranks":[2,2,4,3,2,3,3,2,3,3,3,3]},{"rank":4,"last":4,"name":"Z.ai","share":6.62,"prevShare":7.79,"peak":2,"weeks":12,"series":[9.81,15.83,17.33,16.87,14.55,14.84,10.48,9.52,7.14,6.2,7.79,6.62],"ranks":[4,4,3,2,4,4,4,5,5,4,4,4]},{"rank":5,"last":5,"name":"Google","share":4.47,"prevShare":4.1,"peak":4,"weeks":12,"series":[9.05,6.94,7.05,6.88,5.92,3.17,4.89,9.66,7.39,3.65,4.1,4.47],"ranks":[5,5,5,5,5,5,5,4,4,6,5,5]},{"rank":6,"last":6,"name":"NVIDIA","share":3.59,"prevShare":3.98,"peak":5,"weeks":12,"series":[4.96,4.25,3.29,2.69,2.52,2.61,3.49,2.98,3.68,4.2,3.98,3.59],"ranks":[6,7,8,6,7,6,6,6,6,5,6,6]},{"rank":7,"last":7,"name":"MiniMax","share":0.94,"prevShare":1.09,"peak":7,"weeks":12,"series":[0.79,0.48,0.76,0.32,0.32,0.26,0.28,0.16,1.12,2.17,1.09,0.94],"ranks":[10,11,10,10,11,12,11,12,7,7,7,7]},{"rank":8,"last":9,"name":"Mistral","share":0.7,"prevShare":0.73,"peak":7,"weeks":12,"series":[1.8,2.28,3.3,2.36,1.18,0.58,0.86,0.69,0.81,1.12,0.73,0.7],"ranks":[8,8,7,7,8,10,10,10,8,8,9,8]},{"rank":9,"last":8,"name":"Moonshot","share":0.69,"prevShare":0.86,"peak":6,"weeks":12,"series":[3.55,4.45,3.62,1.79,2.71,1.26,1.13,0.93,0.46,1.08,0.86,0.69],"ranks":[7,6,6,8,6,8,9,9,9,9,8,9]},{"rank":10,"last":10,"name":"Alibaba","share":0.33,"prevShare":0.26,"peak":8,"weeks":12,"series":[1.52,1.59,0.81,0.66,0.86,0.84,1.3,1.11,0.41,0.41,0.26,0.33],"ranks":[9,9,9,9,9,9,8,8,10,10,10,10]},{"rank":11,"last":null,"name":"Xiaomi","share":0.22,"prevShare":0.95,"peak":7,"weeks":5,"series":[0.12,1.59,2.18,2.33,2.66,2.28,1.6,2.08,1.51,1.02,0.95,0.22],"ranks":[12,null,null,null,null,7,7,7,null,null,null,11]},{"rank":12,"last":11,"name":"xAI","share":0.09,"prevShare":0.14,"peak":10,"weeks":11,"series":[0.34,0.56,0.3,0.22,0.43,0.3,0.35,0.22,0.09,0.08,0.14,0.09],"ranks":[11,10,11,11,10,11,null,11,11,11,11,12]}]},{"id":"hosts","label":"Hosts","unit":"pct","denominator":"of tokens (hosts are not team-capped)","rows":[{"rank":1,"last":1,"name":"Azure","share":16.32,"prevShare":16.11,"peak":1,"weeks":12,"series":[4.76,8.41,8.78,7.95,10.06,16.37,15.91,12.09,14.13,15.98,16.11,16.32],"ranks":[5,2,2,3,3,3,1,1,2,1,1,1]},{"rank":2,"last":2,"name":"Bedrock","share":12.78,"prevShare":12.73,"peak":2,"weeks":12,"series":[4.25,7.82,5.64,4.74,5.63,16.5,14.33,10.95,10.4,11.04,12.73,12.78],"ranks":[6,3,6,7,5,2,3,3,3,3,2,2]},{"rank":3,"last":3,"name":"Sference","share":10.45,"prevShare":11.56,"peak":1,"weeks":12,"series":[9.36,17.91,19.24,22.19,20,19.87,14.43,11.5,14.15,11.48,11.56,10.45],"ranks":[2,1,1,1,1,1,2,2,1,2,3,3]},{"rank":4,"last":6,"name":"Melious","share":9.69,"prevShare":5.79,"peak":4,"weeks":6,"series":[0,0,0,0,0,0,1.46,2.54,7.37,6.09,5.79,9.69],"ranks":[null,null,null,null,null,null,15,11,4,6,6,4]},{"rank":5,"last":5,"name":"DeepInfra","share":7.59,"prevShare":7.54,"peak":2,"weeks":12,"series":[2.24,3.31,3.95,5.08,11.74,5.98,4.63,3.69,7.12,7.51,7.54,7.59],"ranks":[9,11,7,5,2,4,9,8,5,4,5,5]},{"rank":6,"last":4,"name":"Vertex (Claude)","share":7.09,"prevShare":7.77,"peak":4,"weeks":9,"series":[5.4,9.87,6.11,5.37,4.35,5.07,7.22,4.58,6.19,6.41,7.77,7.09],"ranks":[4,null,5,null,null,7,4,6,8,5,4,6]},{"rank":7,"last":8,"name":"Anthropic","share":4.33,"prevShare":4.15,"peak":4,"weeks":12,"series":[3.83,5.69,6.53,6.31,4.46,5.75,4.85,2.93,4.75,4.39,4.15,4.33],"ranks":[7,7,4,4,7,5,6,9,10,9,8,7]},{"rank":8,"last":10,"name":"Vertex (Gemini)","share":4.16,"prevShare":3.95,"peak":3,"weeks":8,"series":[8.25,6.63,12.17,10.22,11.72,2.75,3.96,20.42,6.8,2.78,3.95,4.16],"ranks":[3,5,null,null,null,10,10,null,6,12,10,8]},{"rank":9,"last":7,"name":"OpenAI","share":3.74,"prevShare":4.6,"peak":6,"weeks":12,"series":[2.15,3.38,3.63,5.01,4.94,4.94,4.79,3.89,3.63,4.28,4.6,3.74],"ranks":[10,10,9,6,6,8,7,7,11,10,7,9]},{"rank":10,"last":12,"name":"TensorX","share":3.58,"prevShare":3.61,"peak":5,"weeks":12,"series":[2.03,4.95,3.64,2.79,2.62,4,4.75,5.13,5.34,5.94,3.61,3.58],"ranks":[11,8,8,10,8,9,8,5,9,7,12,10]},{"rank":11,"last":9,"name":"DeepSeek","share":3.43,"prevShare":4.13,"peak":2,"weeks":11,"series":[34.1,7.36,8.15,9.92,8.72,5.27,6.82,6.28,6.26,5.28,4.13,3.43],"ranks":[null,4,3,2,4,6,5,4,7,8,9,11]},{"rank":12,"last":11,"name":"NVIDIA","share":3.06,"prevShare":3.63,"peak":8,"weeks":12,"series":[2.88,3.47,2.99,2.4,2.05,2.29,3.11,2.39,3.27,3.87,3.63,3.06],"ranks":[8,9,11,11,10,11,11,12,12,11,11,12]},{"rank":13,"last":16,"name":"Fireworks","share":1.12,"prevShare":0.82,"peak":11,"weeks":12,"series":[1.14,1.5,2.28,1.98,1.97,1.45,2.68,2.31,0.74,1.15,0.82,1.12],"ranks":[13,15,15,13,11,12,12,13,17,16,16,13]},{"rank":14,"last":14,"name":"Mistral","share":0.83,"prevShare":1.14,"peak":12,"weeks":12,"series":[1.04,2.03,2.84,2.07,0.96,0.51,0.76,0.56,0.81,1.49,1.14,0.83],"ranks":[14,13,12,12,14,18,20,18,16,13,14,14]},{"rank":15,"last":15,"name":"Novita","share":0.57,"prevShare":1.03,"peak":9,"weeks":10,"series":[1.33,1.49,2.36,3.29,2.26,1.23,1.32,0.94,0.6,0.9,1.03,0.57],"ranks":[null,null,14,9,9,13,17,16,19,18,15,15]},{"rank":16,"last":18,"name":"Z.ai","share":0.29,"prevShare":0.33,"peak":12,"weeks":12,"series":[1.27,2.62,1.51,1.22,0.95,0.67,0.82,1.06,0.64,0.34,0.33,0.29],"ranks":[12,12,16,15,16,15,19,15,18,19,18,16]},{"rank":17,"last":20,"name":"Alibaba","share":0.14,"prevShare":0.15,"peak":15,"weeks":12,"series":[0.54,1,0.44,0.23,0.38,0.53,0.85,0.52,0.13,0.14,0.15,0.14],"ranks":[15,16,17,16,18,17,18,19,21,21,20,17]},{"rank":18,"last":null,"name":"Xiaomi","share":0.14,"prevShare":0.82,"peak":18,"weeks":2,"series":[0.06,1.54,2.1,2.64,2.17,1.66,1.3,1.4,0.76,0.66,0.82,0.14],"ranks":[21,null,null,null,null,null,null,null,null,null,null,18]},{"rank":19,"last":null,"name":"Inceptron","share":0.11,"prevShare":0.09,"peak":19,"weeks":9,"series":[0.17,0.19,0.06,0.04,0.07,0.08,0.09,0.06,0.06,0.11,0.09,0.11],"ranks":[19,21,22,22,21,23,null,23,null,23,null,19]},{"rank":20,"last":19,"name":"Moonshot","share":0.1,"prevShare":0.23,"peak":17,"weeks":10,"series":[0.35,0.68,0.37,0.2,0.14,0.12,0.11,0.06,0.09,0.3,0.23,0.1],"ranks":[17,17,19,17,null,22,21,null,22,20,19,20]}]}],"vitals":[{"id":"cached","label":"Input read from cache","metric":{"value":78.1,"unit":"pct","denominator":"of input tokens","prev":77.1,"better":"up"},"series":[81.7,76.3,71.9,74.9,73.6,76.3,75.3,74.4,77.1,78.1,77.1,78.1],"note":"Share of input tokens served from a prompt cache."},{"id":"io","label":"Input per output token","metric":{"value":46.3,"unit":"ratio","denominator":"input tokens per output token","prev":46},"series":[39.9,32.9,24.8,26.2,32.7,28,23.4,34.5,32.9,40,46,46.3],"note":"Tokens read for each token written."},{"id":"reason","label":"Output spent thinking","metric":{"value":31.5,"unit":"pct","denominator":"of output tokens","prev":33.9},"series":[39.6,41.3,33.6,16.3,21.8,21.4,36.2,27.9,24.2,36.8,33.9,31.5],"note":"Reasoning tokens as a share of output."},{"id":"tool","label":"Turns ending in a tool call","metric":{"value":18.5,"unit":"pct","denominator":"of responses","prev":20.1},"series":[31.5,22.7,17.8,14.5,20.8,14,16.7,18.7,17.8,21,20.1,18.5],"note":"Responses that asked for a tool instead of answering."},{"id":"ttft","label":"Median first token","metric":{"value":1407,"unit":"ms","denominator":"median, streaming requests","prev":1523,"better":"down"},"series":[679,794,731,919,1426,1555,1489,1949,1529,1602,1523,1407],"note":"Across every streaming request."},{"id":"tps","label":"Median stream speed","metric":{"value":171.4,"unit":"tok_s","denominator":"median, streaming requests","prev":155.6,"better":"up"},"series":[71.8,79,84.9,96.4,106.3,126.7,139.4,163.3,156.5,169.7,155.6,171.4],"note":"Once the model starts talking."},{"id":"policy","label":"Routed by a policy","metric":{"value":35.9,"unit":"pct","denominator":"of tokens","prev":36.8,"better":"up"},"series":[15.7,32.4,39.1,37.2,42.1,40.3,34.8,30,36.4,37.7,36.8,35.9],"note":"Fallback, latency or load-balancing policies."},{"id":"latency_policy","label":"Routed by latency","metric":{"value":19,"unit":"pct","denominator":"of tokens","prev":21.6,"better":"up"},"series":[4.3,8,6.6,6,7.8,18.3,22.1,15.8,19.1,21.4,21.6,19],"note":"Tokens where the gateway picked the quickest host."}],"method":["Aggregated from the Requesty production gateway for the week (Monday to Sunday, UTC), compared with the week before and the ten weeks before that.","We publish shares, rates, ranks and latency only: no volumes, no customer data, no prompt content.","In the model, lab and app charts each team counts for at most 2% of a week's tokens, so one heavy team can't bend the chart.","An entry is named only when at least 20 teams used it and no single team is more than 60% of it.","Latency is measured at the gateway on streaming requests. It includes the host and the network, and reflects the prompts teams sent, not a controlled benchmark.","Bring-your-own-key traffic, internal test traffic, synthetic load and abusive accounts are excluded."],"sources":[{"label":"model share","sql":"scripts/weekly/sql/model_share.sql"},{"label":"app share","sql":"scripts/weekly/sql/app_share.sql"},{"label":"lab share","sql":"scripts/weekly/sql/lab_share.sql"},{"label":"host share errors","sql":"scripts/weekly/sql/host_share_errors.sql"},{"label":"vitals","sql":"scripts/weekly/sql/vitals.sql"},{"label":"breadth","sql":"scripts/weekly/sql/breadth.sql"},{"label":"request settings","sql":"scripts/weekly/sql/request_settings.sql"},{"label":"speed","sql":"scripts/weekly/sql/speed.sql"},{"label":"rescues","sql":"scripts/weekly/sql/rescues.sql"},{"label":"flow","sql":"scripts/weekly/sql/flow.sql"},{"label":"latency dist","sql":"scripts/weekly/sql/latency_dist.sql"},{"label":"host race","sql":"scripts/weekly/sql/host_race.sql"},{"label":"cache ttft","sql":"scripts/weekly/sql/cache_ttft.sql"},{"label":"tools","sql":"scripts/weekly/sql/tools.sql"},{"label":"sessions","sql":"scripts/weekly/sql/sessions.sql"},{"label":"pulse","sql":"scripts/weekly/sql/pulse.sql"},{"label":"host hourly","sql":"scripts/weekly/sql/host_hourly.sql"}],"deep":{"breadth":{"bins":["1","2","3","4","5","6","7","8","9","10","11","12+"],"rows":[{"id":"models","label":"Models","pct":[32.4,19,13.6,9.8,6.7,4,3.2,3.2,1.6,1,0.7,4.9],"prev":[31,18.8,13.4,9.1,7.4,5.3,3.5,2.8,1.2,1.3,0.9,5.5]},{"id":"labs","label":"Labs","pct":[41.4,25.4,13.7,8,4.2,2.3,1.7,1.1,0.4,0.2,0.1,1.6],"prev":[40.2,24.6,15.3,9,3.4,2.1,1.3,1.4,0.6,0.3,0.2,1.5]},{"id":"hosts","label":"Providers","pct":[39.8,26.3,12.2,7.9,4.4,2.1,1.9,1.1,0.9,0.6,0.5,2.4],"prev":[38.7,26.2,13.6,6.9,4.6,2.4,1.7,1.5,0.8,0.9,0.6,2.2]}],"denominator":"of active teams (5+ chat requests in the week)"},"flow":{"apps":[{"name":"OpenAI Python SDK","label":"OpenAI Python SDK","share":21.27,"kind":"sdk"},{"name":"OpenCode","label":"OpenCode","share":13.14,"kind":"agent"},{"name":"Vercel AI SDK","label":"Vercel AI SDK","share":10.06,"kind":"sdk"},{"name":"Claude Code","label":"Claude Code","share":9.82,"kind":"agent"},{"name":"Pi","label":"Pi","share":6.64,"kind":"agent"},{"name":"JS runtime fetch","label":"JS runtime fetch","share":5.49,"kind":"generic"},{"name":"Python HTTP","label":"Python HTTP","share":3.86,"kind":"generic"},{"name":"Anthropic SDK","label":"Anthropic SDK","share":2.5,"kind":"sdk"},{"name":"OpenAI SDK (other)","label":"OpenAI SDK (other)","share":2.36,"kind":"sdk"},{"name":"DeepSeek Harness","label":"DeepSeek Harness","share":2.15,"kind":"agent"},{"name":"Everything else","label":"Everything else","share":22.7,"kind":"generic"}],"labs":[{"name":"DeepSeek","label":"DeepSeek","share":34.07},{"name":"Anthropic","label":"Anthropic","share":24.4},{"name":"OpenAI","label":"OpenAI","share":19.69},{"name":"Z.ai","label":"Z.ai","share":6.62},{"name":"Google","label":"Google","share":4.47},{"name":"NVIDIA","label":"NVIDIA","share":3.59},{"name":"MiniMax","label":"MiniMax","share":0.94},{"name":"Other labs","label":"Other labs","share":6.23}],"hosts":[{"name":"azure","label":"Azure","share":15.28},{"name":"bedrock","label":"Bedrock","share":12.92},{"name":"sference","label":"Sference","share":11.62},{"name":"deepinfra","label":"DeepInfra","share":8.9},{"name":"melious","label":"Melious","share":7.28},{"name":"vertex_claude","label":"Vertex (Claude)","share":6.6},{"name":"anthropic","label":"Anthropic","share":4.84},{"name":"openai","label":"OpenAI","share":4.41},{"name":"vertex_gemini","label":"Vertex (Gemini)","share":4.08},{"name":"deepseek","label":"DeepSeek","share":4.05},{"name":"other hosts","label":"Other hosts","share":20.03}],"al":[{"a":"Claude Code","l":"Anthropic","v":8.21},{"a":"OpenAI Python SDK","l":"OpenAI","v":4.71},{"a":"Everything else","l":"Other labs","v":3.53},{"a":"OpenCode","l":"OpenAI","v":3.95},{"a":"Everything else","l":"DeepSeek","v":9.33},{"a":"OpenAI Python SDK","l":"DeepSeek","v":8.64},{"a":"JS runtime fetch","l":"NVIDIA","v":2.33},{"a":"Vercel AI SDK","l":"OpenAI","v":2.29},{"a":"Anthropic SDK","l":"Anthropic","v":2.31},{"a":"Vercel AI SDK","l":"Anthropic","v":2.36},{"a":"Everything else","l":"OpenAI","v":4.09},{"a":"OpenAI Python SDK","l":"Anthropic","v":4.13},{"a":"Vercel AI SDK","l":"DeepSeek","v":3.98},{"a":"Pi","l":"OpenAI","v":1.66},{"a":"Pi","l":"DeepSeek","v":3.31},{"a":"OpenCode","l":"Z.ai","v":1.69},{"a":"OpenCode","l":"DeepSeek","v":4.08},{"a":"OpenCode","l":"Anthropic","v":2.16},{"a":"OpenAI Python SDK","l":"Z.ai","v":1.9},{"a":"OpenAI Python SDK","l":"Google","v":1.16},{"a":"JS runtime fetch","l":"OpenAI","v":1.01},{"a":"Python HTTP","l":"Google","v":0.91},{"a":"Everything else","l":"Anthropic","v":3.31},{"a":"DeepSeek Harness","l":"DeepSeek","v":1.2},{"a":"Python HTTP","l":"DeepSeek","v":1.57},{"a":"OpenAI SDK (other)","l":"OpenAI","v":0.88},{"a":"Vercel AI SDK","l":"Google","v":0.64},{"a":"Everything else","l":"Google","v":0.87},{"a":"Vercel AI SDK","l":"Z.ai","v":0.65},{"a":"Claude Code","l":"OpenAI","v":0.62},{"a":"Pi","l":"Anthropic","v":0.67},{"a":"DeepSeek Harness","l":"Z.ai","v":0.63},{"a":"JS runtime fetch","l":"DeepSeek","v":0.59},{"a":"JS runtime fetch","l":"Other labs","v":0.86},{"a":"Everything else","l":"MiniMax","v":0.36},{"a":"Pi","l":"Z.ai","v":0.6},{"a":"OpenCode","l":"NVIDIA","v":0.28},{"a":"OpenAI SDK (other)","l":"DeepSeek","v":0.6},{"a":"OpenCode","l":"MiniMax","v":0.29},{"a":"Everything else","l":"NVIDIA","v":0.57},{"a":"Claude Code","l":"DeepSeek","v":0.68},{"a":"Python HTTP","l":"OpenAI","v":0.4},{"a":"Python HTTP","l":"Anthropic","v":0.53},{"a":"OpenCode","l":"Google","v":0.22},{"a":"JS runtime fetch","l":"Anthropic","v":0.49},{"a":"Pi","l":"Google","v":0.21},{"a":"OpenAI SDK (other)","l":"Google","v":0.21},{"a":"Python HTTP","l":"Other labs","v":0.34},{"a":"OpenAI SDK (other)","l":"Anthropic","v":0.22},{"a":"Claude Code","l":"Google","v":0.14},{"a":"OpenCode","l":"Other labs","v":0.47},{"a":"Vercel AI SDK","l":"NVIDIA","v":0.13},{"a":"OpenAI Python SDK","l":"NVIDIA","v":0.11},{"a":"DeepSeek Harness","l":"NVIDIA","v":0.1},{"a":"Everything else","l":"Z.ai","v":0.63},{"a":"OpenAI Python SDK","l":"Other labs","v":0.55},{"a":"Pi","l":"MiniMax","v":0.09},{"a":"OpenAI SDK (other)","l":"Z.ai","v":0.25},{"a":"JS runtime fetch","l":"Z.ai","v":0.11},{"a":"DeepSeek Harness","l":"MiniMax","v":0.09},{"a":"Anthropic SDK","l":"DeepSeek","v":0.09},{"a":"JS runtime fetch","l":"Google","v":0.1},{"a":"OpenAI Python SDK","l":"MiniMax","v":0.08},{"a":"Claude Code","l":"Z.ai","v":0.07},{"a":"OpenAI SDK (other)","l":"Other labs","v":0.19},{"a":"DeepSeek Harness","l":"Other labs","v":0.06},{"a":"Pi","l":"Other labs","v":0.11},{"a":"Python HTTP","l":"Z.ai","v":0.09},{"a":"DeepSeek Harness","l":"OpenAI","v":0.05},{"a":"Claude Code","l":"Other labs","v":0.06}],"lh":[{"l":"Anthropic","h":"bedrock","v":12.92},{"l":"OpenAI","h":"azure","v":15.23},{"l":"Other labs","h":"other hosts","v":5.88},{"l":"DeepSeek","h":"melious","v":6.63},{"l":"DeepSeek","h":"deepinfra","v":8.07},{"l":"NVIDIA","h":"other hosts","v":3.59},{"l":"DeepSeek","h":"other hosts","v":7.43},{"l":"Anthropic","h":"vertex_claude","v":6.6},{"l":"Anthropic","h":"anthropic","v":4.84},{"l":"DeepSeek","h":"deepseek","v":4.05},{"l":"DeepSeek","h":"sference","v":7.89},{"l":"OpenAI","h":"openai","v":4.41},{"l":"Z.ai","h":"sference","v":3.55},{"l":"Google","h":"vertex_gemini","v":4.08},{"l":"Z.ai","h":"deepinfra","v":0.63},{"l":"Z.ai","h":"other hosts","v":1.79},{"l":"MiniMax","h":"other hosts","v":0.94},{"l":"Z.ai","h":"melious","v":0.65},{"l":"Other labs","h":"sference","v":0.18},{"l":"Google","h":"other hosts","v":0.3},{"l":"Google","h":"deepinfra","v":0.08},{"l":"Other labs","h":"deepinfra","v":0.11},{"l":"Other labs","h":"azure","v":0.06},{"l":"OpenAI","h":"other hosts","v":0.05}],"hostKinds":[{"name":"Inference Providers","label":"Inference Providers","share":45.16},{"name":"Hyperscalers","label":"Hyperscalers","share":38.93},{"name":"AI Labs","label":"AI Labs","share":15.91}],"lk":[{"l":"Anthropic","h":"Hyperscalers","v":19.53},{"l":"OpenAI","h":"Hyperscalers","v":15.25},{"l":"Other labs","h":"Inference Providers","v":4.86},{"l":"DeepSeek","h":"Inference Providers","v":30.01},{"l":"NVIDIA","h":"Inference Providers","v":3.59},{"l":"Anthropic","h":"AI Labs","v":4.86},{"l":"DeepSeek","h":"AI Labs","v":4.06},{"l":"OpenAI","h":"AI Labs","v":4.43},{"l":"Z.ai","h":"Inference Providers","v":6.07},{"l":"Google","h":"Hyperscalers","v":4.09},{"l":"Other labs","h":"AI Labs","v":1.31},{"l":"MiniMax","h":"AI Labs","v":0.47},{"l":"MiniMax","h":"Inference Providers","v":0.47},{"l":"Z.ai","h":"AI Labs","v":0.55},{"l":"Google","h":"AI Labs","v":0.23},{"l":"Google","h":"Inference Providers","v":0.15},{"l":"Other labs","h":"Hyperscalers","v":0.06}],"denominator":"of tokens, each team capped at 2%"},"hostForm":[{"host":"azure","label":"Azure","share":16.32,"errs":[3.42,0.65,4.01,0.55,0.61,32.01,0.55,0.83,0.76,3.21,3.17,0.05],"err":0.05,"prevErr":3.17,"ttft":1810,"prevTtft":1872},{"host":"bedrock","label":"Bedrock","share":12.78,"errs":[0.31,0.12,0.43,0.56,0.05,3.52,9,0.64,0.03,0.2,0.24,2.3],"err":2.3,"prevErr":0.24,"ttft":2554,"prevTtft":2757},{"host":"sference","label":"Sference","share":10.45,"errs":[0.35,0.41,6.09,0.76,0.16,3.7,3.67,1.67,0.21,6.21,3.09,0.68],"err":0.68,"prevErr":3.09,"ttft":454,"prevTtft":529},{"host":"melious","label":"Melious","share":9.69,"errs":[null,null,null,null,null,null,0.01,0.07,0.15,0.88,9.94,0.26],"err":0.26,"prevErr":9.94,"ttft":663,"prevTtft":722},{"host":"deepinfra","label":"DeepInfra","share":7.59,"errs":[0.58,0.67,44.53,35.02,5.85,1.09,1.94,2.15,7.5,16.61,0.32,0.08],"err":0.08,"prevErr":0.32,"ttft":1420,"prevTtft":1418},{"host":"vertex_claude","label":"Vertex (Claude)","share":7.09,"errs":[14.28,0.01,0,0.12,0.15,0.08,0.92,0.05,0.05,0.03,0.05,0.02],"err":0.02,"prevErr":0.05,"ttft":1904,"prevTtft":1977},{"host":"anthropic","label":"Anthropic","share":4.33,"errs":[0.05,0.24,0.09,26.09,0.22,0.18,7.91,0.07,0,2.17,0,0.11],"err":0.11,"prevErr":0,"ttft":2066,"prevTtft":2335},{"host":"vertex_gemini","label":"Vertex (Gemini)","share":4.16,"errs":[1.23,0.55,0.3,0.28,0.22,0.23,0.89,0.37,0.34,0.36,0.06,0.07],"err":0.07,"prevErr":0.06,"ttft":2153,"prevTtft":2076},{"host":"openai","label":"OpenAI","share":3.74,"errs":[0.36,0.23,0.05,0.04,0.26,0.13,0.3,0.22,0.06,0.35,0.23,0.16],"err":0.16,"prevErr":0.23,"ttft":2657,"prevTtft":2582},{"host":"tensorx","label":"TensorX","share":3.58,"errs":[9.33,0.45,0.41,1.61,0.59,4.11,1.58,15.92,1.44,2.55,0.62,0.49],"err":0.49,"prevErr":0.62,"ttft":1281,"prevTtft":1492},{"host":"deepseek","label":"DeepSeek","share":3.43,"errs":[0.04,0.09,0.01,0.05,0.08,0,0.01,0.29,0.01,0.03,0.79,0.03],"err":0.03,"prevErr":0.79,"ttft":1127,"prevTtft":1140},{"host":"nvidia","label":"NVIDIA","share":3.06,"errs":[3.65,5.01,5,6.01,6.55,4.46,5.32,7.44,1.51,1.55,1.48,1.03],"err":1.03,"prevErr":1.48,"ttft":1619,"prevTtft":1660},{"host":"fireworks","label":"Fireworks","share":1.12,"errs":[0.34,0.13,0.65,0.42,4.12,2.32,0.32,0.18,0.16,0.03,2.48,0.33],"err":0.33,"prevErr":2.48,"ttft":776,"prevTtft":1189},{"host":"mistral","label":"Mistral","share":0.83,"errs":[0.28,0.17,0.79,0.15,0.17,0.38,1.06,0.75,0.43,0.27,0.11,0.69],"err":0.69,"prevErr":0.11,"ttft":1309,"prevTtft":1165},{"host":"novita","label":"Novita","share":0.57,"errs":[1.87,2.82,0.98,2.73,0.36,1.22,1.78,0.55,3.79,0.01,2.5,0.04],"err":0.04,"prevErr":2.5,"ttft":2905,"prevTtft":3183}],"latency":[{"model":"glm-5.3-flash","lab":"Z.ai","ttft":[246,397,746,1506,3295,14425],"tps":[73.4,115.4,163.5,222.9,283.3],"ttftBins":[[38,0],[39,0.002],[40,0.007],[41,0.037],[42,0.309],[43,0.971],[44,1.911],[45,1.884],[46,2.626],[47,2.831],[48,3.065],[49,3.221],[50,3.878],[51,4.356],[52,4.442],[53,4.458],[54,4.58],[55,4.679],[56,4.623],[57,4.711],[58,4.709],[59,4.572],[60,4.17],[61,3.792],[62,3.401],[63,3.051],[64,2.662],[65,2.505],[66,2.363],[67,2.061],[68,1.897],[69,1.683],[70,1.479],[71,1.332],[72,1.222],[73,1.078],[74,0.916],[75,0.772],[76,0.666],[77,0.567],[78,0.467],[79,0.365],[80,0.256],[81,0.214],[82,0.182],[83,0.154],[84,0.136],[85,0.119],[86,0.097],[87,0.083],[88,0.07],[89,0.059],[90,0.048],[91,0.034],[92,0.034],[93,0.03],[94,0.03],[95,0.022],[96,0.016],[97,0.016],[98,0.011],[99,0.011],[100,0.009],[101,0.006],[102,0.006],[103,0.005],[104,0.006],[105,0.004],[106,0.007],[107,0.003],[108,0.002],[109,0.001],[110,0],[111,0],[112,0.001],[113,0],[114,0.001],[115,0.001],[116,0.001],[117,0.001],[118,0],[120,0],[121,0]],"tpsBins":[[-11,0],[-5,0],[-3,0],[-2,0],[5,0.001],[6,0],[7,0],[9,0],[10,0.001],[11,0.001],[12,0.001],[13,0.002],[14,0.002],[15,0.001],[16,0.003],[17,0.003],[18,0.002],[19,0.005],[20,0.008],[21,0.004],[22,0.01],[23,0.012],[24,0.021],[25,0.033],[26,0.038],[27,0.056],[28,0.085],[29,0.132],[30,0.229],[31,0.409],[32,0.751],[33,1.125],[34,1.592],[35,2.049],[36,2.547],[37,2.875],[38,3.201],[39,3.76],[40,4.707],[41,6.018],[42,8.658],[43,9.047],[44,9.678],[45,9.414],[46,8.799],[47,7.731],[48,6.723],[49,5.06],[50,2.885],[51,1.248],[52,0.553],[53,0.221],[54,0.113],[55,0.07],[56,0.036],[57,0.019],[58,0.013],[59,0.012],[60,0.005],[61,0.005],[62,0.005],[63,0.001],[64,0.001],[65,0.001],[66,0.002],[67,0.001],[68,0.001],[69,0],[70,0.001],[71,0.001],[72,0.001],[74,0.001],[75,0],[76,0.001],[77,0],[80,0.001],[94,0]]},{"model":"deepseek-v4.1-flash","lab":"DeepSeek","ttft":[271,444,834,1442,2512,13446],"tps":[101.7,141.1,193.7,272.4,373.3],"ttftBins":[[35,0],[36,0.008],[37,0.032],[38,0.087],[39,0.206],[40,0.262],[41,0.475],[42,0.705],[43,0.788],[44,0.961],[45,1.214],[46,1.554],[47,2.081],[48,2.551],[49,2.874],[50,3.274],[51,4.067],[52,4.104],[53,4.252],[54,4.411],[55,4.503],[56,4.632],[57,4.848],[58,5.125],[59,5.339],[60,5.495],[61,5.377],[62,4.948],[63,4.391],[64,3.797],[65,3.134],[66,2.515],[67,1.993],[68,1.582],[69,1.254],[70,1.027],[71,0.859],[72,0.724],[73,0.644],[74,0.56],[75,0.503],[76,0.421],[77,0.363],[78,0.304],[79,0.258],[80,0.218],[81,0.184],[82,0.164],[83,0.135],[84,0.114],[85,0.101],[86,0.078],[87,0.064],[88,0.056],[89,0.047],[90,0.041],[91,0.034],[92,0.026],[93,0.025],[94,0.022],[95,0.022],[96,0.018],[97,0.016],[98,0.013],[99,0.011],[100,0.011],[101,0.01],[102,0.009],[103,0.008],[104,0.01],[105,0.012],[106,0.016],[107,0.014],[108,0.003],[109,0.002],[110,0.001],[111,0.001],[112,0.001],[113,0.001],[114,0.002],[115,0.002],[116,0.003],[117,0.002],[118,0.001],[119,0.001],[124,0],[131,0]],"tpsBins":[[-23,0],[-13,0],[-11,0],[-8,0],[-4,0],[0,0],[1,0],[2,0],[5,0],[7,0],[9,0],[10,0],[11,0],[12,0.001],[13,0],[14,0.001],[15,0.001],[16,0.001],[17,0.001],[18,0.002],[19,0.001],[20,0.003],[21,0.004],[22,0.005],[23,0.01],[24,0.012],[25,0.019],[26,0.022],[27,0.034],[28,0.045],[29,0.064],[30,0.088],[31,0.119],[32,0.168],[33,0.23],[34,0.316],[35,0.458],[36,0.664],[37,1.165],[38,2.284],[39,3.652],[40,4.504],[41,5.098],[42,6.072],[43,7.616],[44,9.651],[45,10.187],[46,9.15],[47,8.325],[48,6.975],[49,5.953],[50,5.122],[51,4.172],[52,2.986],[53,1.982],[54,1.189],[55,0.665],[56,0.35],[57,0.191],[58,0.102],[59,0.055],[60,0.034],[61,0.028],[62,0.025],[63,0.021],[64,0.02],[65,0.015],[66,0.015],[67,0.015],[68,0.014],[69,0.012],[70,0.011],[71,0.01],[72,0.009],[73,0.009],[74,0.006],[75,0.005],[76,0.005],[77,0.004],[78,0.004],[79,0.002],[80,0.003],[81,0.002],[82,0.002],[83,0.002],[84,0.002],[85,0.002],[86,0.002],[87,0.002],[88,0.001],[89,0.001],[90,0],[91,0],[92,0.001],[93,0],[94,0],[95,0],[97,0],[107,0],[110,0]]},{"model":"deepseek-v4-flash","lab":"DeepSeek","ttft":[224,416,1140,2584,6188,40567],"tps":[48.4,79.6,127,212.1,313.3],"ttftBins":[[37,0.006],[38,0.232],[39,1.247],[40,1.463],[41,0.714],[42,0.726],[43,0.832],[44,2.241],[45,1.349],[46,1.179],[47,1.71],[48,2.663],[49,3.097],[50,3.038],[51,3.245],[52,3.258],[53,3.044],[54,2.833],[55,2.849],[56,2.764],[57,2.784],[58,2.758],[59,2.687],[60,2.867],[61,2.959],[62,3.834],[63,3.749],[64,3.697],[65,3.773],[66,3.61],[67,3.148],[68,2.592],[69,2.395],[70,2.158],[71,1.99],[72,1.916],[73,1.76],[74,1.591],[75,1.465],[76,1.186],[77,1.065],[78,0.961],[79,0.893],[80,0.741],[81,0.632],[82,0.537],[83,0.446],[84,0.381],[85,0.336],[86,0.312],[87,0.254],[88,0.289],[89,0.268],[90,0.236],[91,0.215],[92,0.18],[93,0.186],[94,0.168],[95,0.133],[96,0.087],[97,0.051],[98,0.036],[99,0.03],[100,0.027],[101,0.021],[102,0.015],[103,0.01],[104,0.007],[105,0.009],[106,0.005],[107,0.006],[108,0.004],[109,0.006],[110,0.008],[111,0.008],[112,0.006],[113,0.007],[114,0.005],[115,0.005],[116,0.001],[117,0],[119,0.001],[120,0.001]],"tpsBins":[[-9,0],[-3,0],[-2,0],[0,0.001],[1,0.002],[2,0.003],[3,0.004],[4,0.003],[5,0.005],[6,0.006],[7,0.008],[8,0.009],[9,0.014],[10,0.016],[11,0.016],[12,0.025],[13,0.036],[14,0.032],[15,0.033],[16,0.038],[17,0.046],[18,0.065],[19,0.085],[20,0.14],[21,0.151],[22,0.213],[23,0.27],[24,0.35],[25,0.412],[26,0.484],[27,0.543],[28,0.681],[29,0.912],[30,1.075],[31,1.321],[32,1.615],[33,2.085],[34,2.542],[35,3.162],[36,3.868],[37,4.658],[38,5.477],[39,6.026],[40,6.519],[41,6.607],[42,6],[43,5.808],[44,5.604],[45,5.24],[46,5.176],[47,4.897],[48,4.332],[49,3.655],[50,3.414],[51,2.685],[52,1.64],[53,0.96],[54,0.553],[55,0.348],[56,0.086],[57,0.011],[58,0.007],[59,0.008],[60,0.003],[61,0.004],[62,0.001],[63,0.002],[64,0],[65,0.002],[66,0],[67,0.001],[70,0],[71,0.001],[72,0.001],[74,0.001],[75,0],[77,0.001]]},{"model":"gpt-6-luna","lab":"OpenAI","ttft":[484,626,1677,4352,8445,24273],"tps":[77.9,113,171.3,338.6,744.4],"ttftBins":[[48,0.002],[49,0.053],[50,0.299],[51,1.219],[52,3.845],[53,6.784],[54,7.413],[55,5.749],[56,4.199],[57,3.322],[58,2.887],[59,2.7],[60,2.57],[61,2.56],[62,2.548],[63,2.586],[64,2.563],[65,2.511],[66,2.616],[67,3.13],[68,3.609],[69,3.438],[70,3.163],[71,2.923],[72,2.995],[73,3.133],[74,3.053],[75,2.659],[76,2.322],[77,2.111],[78,1.911],[79,1.717],[80,1.457],[81,1.181],[82,1.022],[83,0.854],[84,0.691],[85,0.549],[86,0.416],[87,0.324],[88,0.247],[89,0.183],[90,0.143],[91,0.103],[92,0.068],[93,0.051],[94,0.031],[95,0.024],[96,0.02],[97,0.013],[98,0.009],[99,0.007],[100,0.005],[101,0.004],[102,0.002],[103,0.002],[104,0],[105,0],[106,0],[107,0],[108,0],[109,0],[110,0],[111,0],[112,0]],"tpsBins":[[-7,0],[-2,0],[-1,0],[0,0],[1,0],[2,0],[3,0],[4,0.002],[5,0.004],[6,0.004],[7,0.004],[8,0.004],[9,0.006],[10,0.008],[11,0.008],[12,0.009],[13,0.011],[14,0.013],[15,0.018],[16,0.02],[17,0.023],[18,0.029],[19,0.035],[20,0.038],[21,0.049],[22,0.053],[23,0.058],[24,0.068],[25,0.076],[26,0.099],[27,0.12],[28,0.165],[29,0.217],[30,0.309],[31,0.436],[32,0.602],[33,0.822],[34,1.049],[35,1.319],[36,1.935],[37,2.96],[38,3.922],[39,4.657],[40,5.479],[41,6.108],[42,6.912],[43,7.474],[44,6.991],[45,5.678],[46,4.806],[47,4.137],[48,3.508],[49,3.077],[50,2.782],[51,2.536],[52,2.39],[53,2.303],[54,2.135],[55,1.976],[56,1.808],[57,1.66],[58,1.52],[59,1.362],[60,1.139],[61,0.951],[62,0.82],[63,0.671],[64,0.518],[65,0.44],[66,0.345],[67,0.265],[68,0.232],[69,0.177],[70,0.143],[71,0.116],[72,0.097],[73,0.069],[74,0.052],[75,0.041],[76,0.035],[77,0.025],[78,0.018],[79,0.013],[80,0.011],[81,0.009],[82,0.005],[83,0.005],[84,0.003],[85,0.001],[86,0.001],[87,0.001],[88,0],[89,0.001],[90,0],[91,0],[94,0],[98,0]]},{"model":"claude-haiku-5-5","lab":"Anthropic","ttft":[845,1149,1755,2845,4720,21831],"tps":[200.9,237.1,291,454.3,945.3],"ttftBins":[[45,0.001],[46,0.039],[47,0.171],[48,0.418],[49,0.596],[50,0.713],[51,0.65],[52,0.623],[53,0.603],[54,0.631],[55,0.821],[56,1.169],[57,1.787],[58,4.064],[59,6.203],[60,5.401],[61,5.609],[62,6.235],[63,7.139],[64,8.022],[65,7.61],[66,5.975],[67,5.114],[68,5.029],[69,4.307],[70,3.619],[71,3.575],[72,2.84],[73,1.99],[74,1.499],[75,1.174],[76,0.97],[77,0.801],[78,0.689],[79,0.619],[80,0.528],[81,0.46],[82,0.376],[83,0.307],[84,0.247],[85,0.223],[86,0.195],[87,0.166],[88,0.142],[89,0.131],[90,0.112],[91,0.093],[92,0.08],[93,0.064],[94,0.052],[95,0.042],[96,0.025],[97,0.016],[98,0.01],[99,0.007],[100,0.005],[101,0.004],[102,0.002],[103,0.001],[104,0.001],[105,0],[106,0],[108,0],[109,0],[110,0],[111,0],[112,0],[113,0]],"tpsBins":[[15,0],[16,0],[18,0],[19,0],[21,0],[22,0],[23,0.001],[24,0.001],[25,0.001],[26,0.001],[27,0.003],[28,0.005],[29,0.002],[30,0.002],[31,0.006],[32,0.005],[33,0.012],[34,0.013],[35,0.018],[36,0.025],[37,0.031],[38,0.037],[39,0.058],[40,0.096],[41,0.208],[42,0.41],[43,0.88],[44,1.614],[45,5.993],[46,9.843],[47,12.985],[48,14.435],[49,10.822],[50,7.887],[51,5.465],[52,3.7],[53,2.501],[54,2.306],[55,2.414],[56,2.472],[57,2.467],[58,2.247],[59,1.92],[60,1.595],[61,1.324],[62,1.039],[63,0.795],[64,0.643],[65,0.504],[66,0.409],[67,0.345],[68,0.286],[69,0.249],[70,0.221],[71,0.208],[72,0.176],[73,0.178],[74,0.152],[75,0.141],[76,0.122],[77,0.109],[78,0.105],[79,0.09],[80,0.072],[81,0.058],[82,0.056],[83,0.039],[84,0.039],[85,0.034],[86,0.027],[87,0.021],[88,0.018],[89,0.015],[90,0.013],[91,0.009],[92,0.006],[93,0.003],[94,0.003],[95,0.004],[96,0.001],[97,0.001],[98,0.001],[100,0]]},{"model":"nemotron-3-ultra-550b-a55b","lab":"NVIDIA","ttft":[1008,1331,1868,2764,4706,27634],"tps":[28.3,42.3,62.5,89.4,118.2],"ttftBins":[[49,0.001],[50,0.044],[51,0.213],[52,0.193],[53,0.114],[54,0.353],[55,0.694],[56,0.779],[57,1.089],[58,2.625],[59,3.614],[60,5.211],[61,6.524],[62,7.622],[63,8.39],[64,8.796],[65,8.739],[66,7.995],[67,7.084],[68,5.821],[69,4.588],[70,3.596],[71,2.767],[72,2.274],[73,1.784],[74,1.491],[75,1.183],[76,0.92],[77,0.716],[78,0.583],[79,0.537],[80,0.471],[81,0.402],[82,0.371],[83,0.33],[84,0.282],[85,0.236],[86,0.218],[87,0.215],[88,0.172],[89,0.195],[90,0.158],[91,0.151],[92,0.126],[93,0.092],[94,0.061],[95,0.04],[96,0.044],[97,0.039],[98,0.031],[99,0.007],[100,0.002],[102,0.001],[103,0.003],[104,0.003],[105,0.001],[106,0.002],[107,0.001],[108,0.001],[110,0.004],[111,0.001],[117,0.001],[118,0.001],[119,0.001],[120,0.001],[121,0.001]],"tpsBins":[[-1,0.001],[2,0.003],[3,0.005],[4,0.014],[5,0.013],[6,0.02],[7,0.019],[8,0.034],[9,0.05],[10,0.064],[11,0.09],[12,0.081],[13,0.076],[14,0.082],[15,0.121],[16,0.121],[17,0.164],[18,0.17],[19,0.17],[20,0.232],[21,0.291],[22,0.42],[23,0.521],[24,0.745],[25,0.917],[26,1.346],[27,1.768],[28,2.391],[29,3.081],[30,3.922],[31,5.017],[32,5.992],[33,6.95],[34,7.619],[35,8.126],[36,8.293],[37,8.225],[38,7.654],[39,6.982],[40,5.95],[41,4.816],[42,3.336],[43,2.209],[44,1.134],[45,0.45],[46,0.156],[47,0.049],[48,0.019],[49,0.009],[50,0.009],[51,0.015],[52,0.012],[53,0.017],[54,0.017],[55,0.001],[56,0.001],[57,0.001],[59,0.001],[66,0.002],[67,0.001],[72,0.001],[78,0.001]]},{"model":"claude-sonnet-5","lab":"Anthropic","ttft":[855,1305,2323,4603,9611,62947],"tps":[82.5,108.3,148.8,262.5,553.4],"ttftBins":[[48,0.007],[49,0.034],[50,0.197],[51,0.534],[52,0.887],[53,1.164],[54,1.295],[55,1.404],[56,1.541],[57,1.643],[58,2.181],[59,2.812],[60,4.223],[61,5.353],[62,5.567],[63,5.506],[64,4.93],[65,4.74],[66,4.487],[67,4.772],[68,5.076],[69,4.383],[70,3.775],[71,3.618],[72,3.842],[73,3.876],[74,3.121],[75,2.609],[76,2.109],[77,1.76],[78,1.658],[79,1.35],[80,1.136],[81,1.068],[82,0.886],[83,0.737],[84,0.73],[85,0.601],[86,0.535],[87,0.484],[88,0.402],[89,0.375],[90,0.341],[91,0.278],[92,0.274],[93,0.263],[94,0.226],[95,0.213],[96,0.172],[97,0.162],[98,0.149],[99,0.123],[100,0.087],[101,0.097],[102,0.05],[103,0.034],[104,0.041],[105,0.022],[106,0.023],[107,0.01],[108,0.004],[109,0.009],[110,0.002],[111,0.004],[112,0.004],[113,0.001],[116,0.001]],"tpsBins":[[9,0.001],[15,0.001],[19,0.001],[21,0.004],[22,0.005],[23,0.009],[24,0.025],[25,0.048],[26,0.071],[27,0.186],[28,0.247],[29,0.144],[30,0.043],[31,0.027],[32,0.039],[33,0.06],[34,0.217],[35,1.175],[36,2.779],[37,3.737],[38,4.006],[39,6.402],[40,8.488],[41,8.829],[42,9.455],[43,8.398],[44,6.555],[45,5.108],[46,4.073],[47,3.694],[48,3.025],[49,2.729],[50,2.65],[51,2.351],[52,2.087],[53,1.864],[54,1.681],[55,1.443],[56,1.277],[57,0.995],[58,0.928],[59,0.713],[60,0.639],[61,0.492],[62,0.388],[63,0.332],[64,0.291],[65,0.279],[66,0.217],[67,0.221],[68,0.168],[69,0.172],[70,0.146],[71,0.116],[72,0.132],[73,0.118],[74,0.103],[75,0.091],[76,0.103],[77,0.083],[78,0.052],[79,0.056],[80,0.037],[81,0.035],[82,0.031],[83,0.027],[84,0.025],[85,0.013],[86,0.019],[87,0.013],[88,0.009],[89,0.011],[90,0.003],[92,0.003],[93,0.001],[94,0.003]]},{"model":"gpt-5.6-luna","lab":"OpenAI","ttft":[956,1426,2463,4113,6285,19412],"tps":[49.4,119.1,173.7,262.3,467.6],"ttftBins":[[50,0.001],[51,0.005],[52,0.061],[53,0.329],[54,0.923],[55,1.323],[56,1.418],[57,1.708],[58,2.3],[59,3.489],[60,4.382],[61,4.298],[62,4.379],[63,4.759],[64,5.081],[65,5.098],[66,5.616],[67,5.788],[68,5.67],[69,5.227],[70,5.137],[71,6.25],[72,5.68],[73,4.762],[74,3.882],[75,2.509],[76,1.844],[77,1.318],[78,1.177],[79,1.019],[80,0.791],[81,0.799],[82,0.72],[83,0.551],[84,0.44],[85,0.337],[86,0.266],[87,0.205],[88,0.144],[89,0.119],[90,0.074],[91,0.046],[92,0.028],[93,0.013],[94,0.01],[95,0.005],[96,0.004],[97,0.009],[98,0.001],[99,0],[100,0.001],[105,0],[106,0],[107,0.001],[110,0],[111,0]],"tpsBins":[[7,0],[9,0],[10,0],[11,0],[12,0],[13,0],[14,0.002],[15,0],[16,0.004],[17,0.007],[18,0.014],[19,0.026],[20,0.054],[21,0.109],[22,0.187],[23,0.174],[24,0.171],[25,0.221],[26,0.217],[27,0.334],[28,0.605],[29,0.864],[30,1.312],[31,1.99],[32,2.157],[33,1.74],[34,1.312],[35,1.139],[36,1.256],[37,1.462],[38,1.783],[39,2.325],[40,3.267],[41,4.844],[42,7.007],[43,8.445],[44,8.69],[45,8.195],[46,7.46],[47,5.793],[48,4.592],[49,3.765],[50,3.019],[51,2.519],[52,2.156],[53,1.965],[54,1.812],[55,1.546],[56,1.221],[57,0.943],[58,0.683],[59,0.481],[60,0.369],[61,0.273],[62,0.221],[63,0.176],[64,0.178],[65,0.149],[66,0.128],[67,0.102],[68,0.102],[69,0.085],[70,0.077],[71,0.056],[72,0.037],[73,0.041],[74,0.031],[75,0.025],[76,0.025],[77,0.021],[78,0.014],[79,0.007],[80,0.004],[81,0.003],[82,0.002],[83,0.001],[86,0],[92,0],[95,0]]},{"model":"gpt-6.1-terra","lab":"OpenAI","ttft":[1207,1756,2711,4319,6747,15893],"tps":[70.8,108.5,164.5,264.6,558.1],"ttftBins":[[51,0.001],[52,0.004],[53,0.024],[54,0.109],[55,0.239],[56,0.34],[57,0.591],[58,1.16],[59,2.071],[60,3.041],[61,3.842],[62,4.299],[63,4.63],[64,5.197],[65,5.925],[66,6.588],[67,7.166],[68,7.108],[69,6.767],[70,6.282],[71,5.795],[72,5.322],[73,4.676],[74,3.894],[75,3.288],[76,2.697],[77,2.089],[78,1.65],[79,1.253],[80,1.014],[81,0.81],[82,0.638],[83,0.483],[84,0.307],[85,0.211],[86,0.154],[87,0.105],[88,0.071],[89,0.054],[90,0.046],[91,0.023],[92,0.01],[93,0.011],[94,0.005],[95,0.002],[96,0.002],[97,0.002],[98,0.002],[99,0.001],[100,0.001],[101,0.001],[119,0.001]],"tpsBins":[[5,0.002],[6,0.001],[7,0.005],[8,0.003],[9,0.006],[10,0.004],[11,0.005],[12,0.023],[13,0.052],[14,0.057],[15,0.055],[16,0.056],[17,0.044],[18,0.059],[19,0.056],[20,0.05],[21,0.046],[22,0.069],[23,0.075],[24,0.091],[25,0.117],[26,0.131],[27,0.215],[28,0.277],[29,0.394],[30,0.489],[31,0.57],[32,0.71],[33,1.049],[34,1.51],[35,1.691],[36,2.085],[37,2.684],[38,4.213],[39,4.429],[40,5.329],[41,6.539],[42,7.119],[43,7.355],[44,7.131],[45,6.837],[46,6.088],[47,5.228],[48,4.191],[49,3.177],[50,2.433],[51,2.147],[52,1.867],[53,1.762],[54,1.589],[55,1.46],[56,1.185],[57,1.097],[58,1.036],[59,0.919],[60,0.758],[61,0.637],[62,0.574],[63,0.532],[64,0.397],[65,0.316],[66,0.23],[67,0.197],[68,0.173],[69,0.126],[70,0.077],[71,0.053],[72,0.036],[73,0.028],[74,0.018],[75,0.012],[76,0.007],[77,0.008],[78,0.005],[79,0.001],[80,0.001],[81,0.002],[82,0.001]]},{"model":"claude-sonnet-5-5","lab":"Anthropic","ttft":[811,1476,2727,4892,8977,57486],"tps":[152.4,194.4,252.5,377.4,669.2],"ttftBins":[[48,0],[49,0.017],[50,0.099],[51,0.326],[52,0.579],[53,0.891],[54,1.447],[55,1.767],[56,2.117],[57,2.308],[58,2.625],[59,2.736],[60,2.759],[61,2.933],[62,3.115],[63,3.368],[64,3.796],[65,4.492],[66,5.122],[67,5.563],[68,5.521],[69,5.566],[70,5.311],[71,4.861],[72,4.448],[73,4.046],[74,3.702],[75,3.357],[76,2.893],[77,2.309],[78,1.825],[79,1.435],[80,1.24],[81,0.979],[82,0.829],[83,0.731],[84,0.607],[85,0.555],[86,0.473],[87,0.428],[88,0.363],[89,0.312],[90,0.288],[91,0.253],[92,0.209],[93,0.197],[94,0.177],[95,0.158],[96,0.137],[97,0.12],[98,0.099],[99,0.102],[100,0.082],[101,0.068],[102,0.056],[103,0.042],[104,0.04],[105,0.034],[106,0.024],[107,0.021],[108,0.012],[109,0.014],[110,0.007],[111,0.005],[112,0.002],[113,0.002],[114,0],[115,0],[116,0],[117,0]],"tpsBins":[[21,0.001],[23,0.001],[24,0.003],[25,0.002],[26,0.001],[27,0.004],[28,0.003],[29,0.006],[30,0.005],[31,0.005],[32,0.01],[33,0.012],[34,0.023],[35,0.037],[36,0.069],[37,0.115],[38,0.21],[39,0.329],[40,0.671],[41,1.559],[42,3.5],[43,5.39],[44,6.632],[45,8.494],[46,10.823],[47,11.621],[48,9.519],[49,7.301],[50,5.916],[51,4.856],[52,4.055],[53,3.205],[54,2.613],[55,2.07],[56,1.718],[57,1.412],[58,1.174],[59,0.965],[60,0.8],[61,0.707],[62,0.602],[63,0.481],[64,0.434],[65,0.345],[66,0.318],[67,0.272],[68,0.255],[69,0.204],[70,0.163],[71,0.156],[72,0.14],[73,0.118],[74,0.101],[75,0.093],[76,0.085],[77,0.062],[78,0.047],[79,0.05],[80,0.043],[81,0.032],[82,0.033],[83,0.024],[84,0.016],[85,0.015],[86,0.02],[87,0.008],[88,0.009],[89,0.01],[90,0.004],[91,0.006],[92,0.004],[93,0.003],[94,0.004],[95,0.002],[96,0.001],[97,0.001],[98,0.001],[99,0.001],[102,0.001]]},{"model":"gpt-6.1-sol","lab":"OpenAI","ttft":[1489,2103,3510,6479,11885,27761],"tps":[45.3,59.3,79.7,122.4,294],"ttftBins":[[51,0],[52,0.001],[53,0.001],[54,0.001],[55,0.001],[56,0.004],[57,0.031],[58,0.168],[59,0.526],[60,1.307],[61,2.506],[62,3.62],[63,4.221],[64,4.764],[65,5.339],[66,5.555],[67,5.69],[68,5.697],[69,5.636],[70,5.431],[71,5.371],[72,5.092],[73,4.752],[74,4.425],[75,4],[76,3.606],[77,3.29],[78,2.98],[79,2.727],[80,2.28],[81,1.894],[82,1.644],[83,1.6],[84,1.613],[85,1.434],[86,0.971],[87,0.543],[88,0.317],[89,0.239],[90,0.163],[91,0.127],[92,0.093],[93,0.07],[94,0.061],[95,0.053],[96,0.039],[97,0.029],[98,0.025],[99,0.021],[100,0.012],[101,0.01],[102,0.006],[103,0.003],[104,0.003],[105,0.002],[106,0.001],[107,0.001],[108,0.001],[109,0.001],[110,0],[112,0.001],[113,0],[114,0]],"tpsBins":[[-8,0],[-4,0],[0,0],[2,0.001],[4,0.001],[5,0.001],[6,0.001],[7,0.001],[8,0.002],[9,0.002],[10,0.001],[11,0.005],[12,0.004],[13,0.008],[14,0.015],[15,0.02],[16,0.03],[17,0.051],[18,0.055],[19,0.066],[20,0.077],[21,0.091],[22,0.096],[23,0.118],[24,0.16],[25,0.198],[26,0.266],[27,0.38],[28,0.512],[29,0.751],[30,1.216],[31,2.05],[32,3.299],[33,5.002],[34,6.809],[35,8.397],[36,9.556],[37,10.486],[38,9.609],[39,7.404],[40,5.182],[41,3.987],[42,3.222],[43,2.708],[44,2.161],[45,1.753],[46,1.464],[47,1.254],[48,1.139],[49,1.034],[50,0.993],[51,0.894],[52,0.831],[53,0.761],[54,0.684],[55,0.631],[56,0.606],[57,0.542],[58,0.483],[59,0.489],[60,0.449],[61,0.395],[62,0.307],[63,0.26],[64,0.219],[65,0.183],[66,0.152],[67,0.129],[68,0.102],[69,0.072],[70,0.05],[71,0.033],[72,0.025],[73,0.016],[74,0.012],[75,0.009],[76,0.01],[77,0.005],[78,0.005],[79,0.004],[80,0.001],[81,0],[82,0.001],[83,0],[84,0],[85,0],[89,0.001],[90,0]]},{"model":"claude-opus-5-5","lab":"Anthropic","ttft":[1385,2347,3751,6008,10405,47028],"tps":[112.7,139.5,175.1,279.7,515.6],"ttftBins":[[49,0],[50,0.002],[51,0.012],[52,0.037],[53,0.137],[54,0.291],[55,0.344],[56,0.381],[57,0.513],[58,0.746],[59,1.133],[60,1.796],[61,2.342],[62,2.749],[63,3.047],[64,3.183],[65,3.248],[66,3.403],[67,4.311],[68,5.667],[69,6.491],[70,6.842],[71,6.976],[72,7.19],[73,6.142],[74,5.161],[75,4.887],[76,4.202],[77,3.36],[78,2.649],[79,2.115],[80,1.768],[81,1.428],[82,1.196],[83,0.988],[84,0.825],[85,0.754],[86,0.66],[87,0.514],[88,0.421],[89,0.348],[90,0.264],[91,0.224],[92,0.174],[93,0.156],[94,0.133],[95,0.116],[96,0.095],[97,0.093],[98,0.083],[99,0.07],[100,0.071],[101,0.055],[102,0.047],[103,0.041],[104,0.026],[105,0.025],[106,0.026],[107,0.016],[108,0.014],[109,0.006],[110,0.001],[111,0.002],[112,0.001],[113,0.001],[114,0.001],[115,0]],"tpsBins":[[10,0],[16,0.001],[18,0.001],[19,0.001],[20,0.002],[21,0.002],[22,0.001],[23,0.004],[24,0.004],[25,0.003],[26,0.003],[27,0.006],[28,0.007],[29,0.006],[30,0.013],[31,0.022],[32,0.028],[33,0.035],[34,0.049],[35,0.062],[36,0.1],[37,0.2],[38,0.445],[39,2.781],[40,5.98],[41,6.859],[42,9.65],[43,13.502],[44,11.46],[45,7.812],[46,6.448],[47,5.417],[48,4.354],[49,3.79],[50,3.281],[51,2.818],[52,2.429],[53,2],[54,1.656],[55,1.38],[56,1.173],[57,0.955],[58,0.816],[59,0.678],[60,0.564],[61,0.456],[62,0.404],[63,0.346],[64,0.268],[65,0.224],[66,0.195],[67,0.174],[68,0.138],[69,0.139],[70,0.106],[71,0.101],[72,0.091],[73,0.087],[74,0.067],[75,0.061],[76,0.055],[77,0.051],[78,0.043],[79,0.04],[80,0.029],[81,0.025],[82,0.021],[83,0.021],[84,0.014],[85,0.01],[86,0.012],[87,0.005],[88,0.004],[89,0.003],[90,0.003],[91,0.001],[92,0.001],[93,0.001],[94,0.001],[95,0.001],[96,0],[97,0.001],[98,0],[100,0.001]]}],"race":[{"model":"deepseek-v4.1-flash","lab":"DeepSeek","hosts":[{"host":"sference","label":"Sference","share":19.5,"err":0.4,"ttft":[238,375,674,1298,6489],"tps":[259.9,419.8]},{"host":"fireworks","label":"Fireworks","share":2.58,"err":0.15,"ttft":[410,565,852,1941,10172],"tps":[188,290.9]},{"host":"melious","label":"Melious","share":31.2,"err":0.34,"ttft":[447,670,1070,2071,20567],"tps":[151.4,279.6]},{"host":"together","label":"Together","share":3.39,"err":0.04,"ttft":[456,688,1181,2406,18549],"tps":[305.4,532.1]},{"host":"deepseek","label":"DeepSeek","share":8.66,"err":0.02,"ttft":[924,1115,1335,1593,2354],"tps":[206.1,294.9]},{"host":"runware","label":"Runware","share":7.23,"err":0.12,"ttft":[947,1352,2016,3411,12942],"tps":[253.7,515.3]},{"host":"deepinfra","label":"DeepInfra","share":22.54,"err":0.07,"ttft":[1187,1633,2239,3240,8624],"tps":[144.4,234.8]},{"host":"tensorx","label":"TensorX","share":4.64,"err":0.36,"ttft":[794,1714,4607,10318,37441],"tps":[373.3,549.4]}]},{"model":"claude-opus-5-5","lab":"Anthropic","hosts":[{"host":"anthropic","label":"Anthropic","share":14.58,"err":0.05,"ttft":[1778,2948,5167,10749,34442],"tps":[192.7,536.2]},{"host":"vertex_claude","label":"Vertex (Claude)","share":21.46,"err":0,"ttft":[2151,3001,4173,8758,30580],"tps":[166.7,439.2]},{"host":"bedrock","label":"Bedrock","share":63.96,"err":0.17,"ttft":[2814,4400,6561,10839,60559],"tps":[174.7,560.5]}]},{"model":"gpt-6-luna","lab":"OpenAI","hosts":[{"host":"azure","label":"Azure","share":89.17,"err":0.04,"ttft":[599,1582,4271,8219,23454],"tps":[179.4,782.5]},{"host":"openai","label":"OpenAI","share":10.83,"err":0.36,"ttft":[1158,2355,5366,10914,33751],"tps":[112.2,428]}]},{"model":"claude-sonnet-5-5","lab":"Anthropic","hosts":[{"host":"anthropic","label":"Anthropic","share":15.43,"err":0.1,"ttft":[961,2111,3571,7688,60173],"tps":[247.4,665.6]},{"host":"vertex_claude","label":"Vertex (Claude)","share":49.45,"err":0,"ttft":[1358,2352,4301,8610,50529],"tps":[241.9,625.9]},{"host":"bedrock","label":"Bedrock","share":35.12,"err":17.75,"ttft":[2149,3623,5773,9833,62743],"tps":[270.9,723]}]},{"model":"gpt-6.1-sol","lab":"OpenAI","hosts":[{"host":"openai","label":"OpenAI","share":13.43,"err":0.6,"ttft":[2004,3400,7099,14104,43690],"tps":[97.1,462.2]},{"host":"azure","label":"Azure","share":86.57,"err":0.17,"ttft":[2123,3528,6388,11453,25063],"tps":[78.2,258.2]}]},{"model":"glm-5.3-flash","lab":"Z.ai","hosts":[{"host":"sference","label":"Sference","share":54.15,"err":0.04,"ttft":[321,569,1070,2117,8190],"tps":[204.6,310.3]},{"host":"melious","label":"Melious","share":12.21,"err":0.05,"ttft":[334,617,1032,2497,17082],"tps":[144.2,248.3]},{"host":"tensorx","label":"TensorX","share":6.06,"err":0.36,"ttft":[416,711,2147,5322,23427],"tps":[179.6,314.1]},{"host":"fireworks","label":"Fireworks","share":7.61,"err":0.64,"ttft":[887,1222,1826,3064,10439],"tps":[168.1,220.2]},{"host":"deepinfra","label":"DeepInfra","share":10.69,"err":0.01,"ttft":[989,1372,2058,3121,9655],"tps":[96.7,189.6]},{"host":"runware","label":"Runware","share":2.76,"err":0.1,"ttft":[1045,1585,2209,2868,13969],"tps":[285.6,440.4]},{"host":"zai","label":"Z.ai","share":5.31,"err":0,"ttft":[2082,2965,4349,6611,24410],"tps":[60.1,96.2]}]},{"model":"deepseek-v4-flash","lab":"DeepSeek","hosts":[{"host":"deepinfra","label":"DeepInfra","share":14.67,"err":0.02,"ttft":[346,602,1367,2339,9081],"tps":[97.7,178.3]},{"host":"melious","label":"Melious","share":9.47,"err":0.17,"ttft":[387,642,1102,2148,18124],"tps":[154,272.9]},{"host":"sference","label":"Sference","share":42.6,"err":1.51,"ttft":[322,1072,2788,7048,42787],"tps":[179.1,371.9]},{"host":"scaleway","label":"Scaleway","share":1.99,"err":0.28,"ttft":[542,1147,2403,5483,12954],"tps":[206,381.2]},{"host":"deepseek","label":"DeepSeek","share":4.86,"err":0.04,"ttft":[1029,1283,1559,1855,2731],"tps":[207.8,321.1]},{"host":"tensorx","label":"TensorX","share":24.56,"err":0.53,"ttft":[1126,2278,4931,10735,75223],"tps":[89.2,143.1]}]},{"model":"claude-haiku-5-5","lab":"Anthropic","hosts":[{"host":"vertex_claude","label":"Vertex (Claude)","share":32.13,"err":0,"ttft":[796,1250,1906,3636,13129],"tps":[346.6,852.4]},{"host":"anthropic","label":"Anthropic","share":20.24,"err":0.15,"ttft":[1561,1897,2647,4491,21652],"tps":[293.8,989.4]},{"host":"bedrock","label":"Bedrock","share":47.63,"err":0.04,"ttft":[1011,1914,3472,5193,27953],"tps":[235.4,962.1]}]},{"model":"claude-sonnet-5","lab":"Anthropic","hosts":[{"host":"anthropic","label":"Anthropic","share":15.34,"err":0.03,"ttft":[1040,2096,3065,5742,35216],"tps":[148.8,392]},{"host":"vertex_claude","label":"Vertex (Claude)","share":20.09,"err":0.15,"ttft":[894,2207,5248,16477,85666],"tps":[165.2,569.4]},{"host":"bedrock","label":"Bedrock","share":64.57,"err":0.01,"ttft":[1495,2645,5161,9600,62229],"tps":[138.9,644.6]}]},{"model":"gpt-6.1-terra","lab":"OpenAI","hosts":[{"host":"azure","label":"Azure","share":83.08,"err":0.07,"ttft":[1784,2660,4048,6055,12541],"tps":[175.4,558.3]},{"host":"openai","label":"OpenAI","share":16.92,"err":0.38,"ttft":[1651,3074,5689,10374,22865],"tps":[104.1,557.1]}]},{"model":"gpt-5.6-luna","lab":"OpenAI","hosts":[{"host":"openai","label":"OpenAI","share":56.75,"err":0.43,"ttft":[1427,2399,4077,6032,18675],"tps":[139,261.8]},{"host":"azure","label":"Azure","share":43.12,"err":0.03,"ttft":[1422,2510,4154,6782,20238],"tps":[244.6,655.8]}]},{"model":"glm-5.3","lab":"Z.ai","hosts":[{"host":"sference","label":"Sference","share":67.29,"err":0.06,"ttft":[213,475,970,1926,9622],"tps":[159.6,204.3]},{"host":"melious","label":"Melious","share":9.86,"err":0.15,"ttft":[410,881,1820,8675,34347],"tps":[91.7,171]},{"host":"deepinfra","label":"DeepInfra","share":4.22,"err":0.05,"ttft":[796,1184,1795,3768,20254],"tps":[128.7,221.8]},{"host":"fireworks","label":"Fireworks","share":1.63,"err":0.34,"ttft":[838,1323,2974,5352,78200],"tps":[117.6,145.7]},{"host":"tensorx","label":"TensorX","share":11.67,"err":0.73,"ttft":[643,1628,3945,9143,46693],"tps":[132.7,265.4]},{"host":"zai","label":"Z.ai","share":3.07,"err":0,"ttft":[1567,2511,3649,5057,11790],"tps":[83.4,117.2]},{"host":"lyceum","label":"Lyceum","share":1.74,"err":0.14,"ttft":[1081,2527,4977,9352,26574],"tps":[139.5,267.6]}]}],"cache":[{"model":"deepseek-v4.1-flash","lab":"DeepSeek","cold":[[8,350,498,745],[9,224,334,538],[10,231,323,506],[11,373,437,608],[12,238,472,839],[13,470,742,1281],[14,601,927,1533],[15,990,1365,1907],[16,1798,2677,6496],[17,3592,5333,10358],[18,7783,11157,17755],[19,16028,21607,44973]],"warm":[[8,189,373,653],[9,128,194,427],[10,356,397,486],[11,152,436,661],[12,144,218,527],[13,179,425,882],[14,306,647,1064],[15,361,723,1118],[16,470,847,1349],[17,621,1072,1741],[18,1107,1756,2684],[19,1708,2645,4265]]},{"model":"gpt-6-luna","lab":"OpenAI","cold":[[8,579,1435,2828],[9,471,540,666],[10,535,760,1105],[11,559,761,1367],[12,859,1334,2453],[13,1334,1947,3417],[14,1426,3106,5875],[15,2573,4731,8618],[16,3606,6699,11387],[17,4120,7196,11221],[18,5331,7575,10865],[19,9327,12682,18242]],"warm":[[10,897,1648,2316],[11,1366,3118,9199],[12,1278,1947,2923],[13,1332,2483,4014],[14,1869,3642,6421],[15,1626,3465,6064],[16,1995,4232,7600],[17,2492,4557,8808],[18,3239,5036,8394],[19,5496,8162,11966]]},{"model":"claude-haiku-5-5","lab":"Anthropic","cold":[[8,914,1520,2328],[9,318,704,1495],[10,873,1033,1590],[11,1536,1907,2722],[12,1478,1966,3193],[13,996,1396,2238],[14,1216,1932,2855],[15,1355,2072,3444],[16,2369,3670,4664],[17,3585,4740,6067],[18,4566,7508,11045],[19,8105,19376,28666]],"warm":[[9,918,973,1043],[10,867,940,1037],[11,814,943,1173],[12,645,1069,1541],[13,965,1333,2149],[14,1215,1860,3397],[15,1281,2280,4398],[16,1463,2715,4941],[17,1424,2619,3693],[18,1535,2546,3743],[19,1472,2587,4117]]},{"model":"glm-5.3-flash","lab":"Z.ai","cold":[[8,274,524,885],[9,374,475,909],[10,387,492,936],[11,383,518,1010],[12,341,648,1122],[13,500,801,1415],[14,984,1342,2166],[15,1632,2258,3797],[16,3052,4064,6936],[17,5735,7445,11294],[18,12985,17224,22868]],"warm":[[8,195,296,838],[9,266,391,553],[10,366,444,744],[11,218,353,563],[12,240,424,733],[13,267,531,930],[14,334,652,1149],[15,482,887,1615],[16,511,912,1586],[17,656,1141,2043],[18,988,1837,3082],[19,2548,3265,4414]]},{"model":"claude-opus-5-5","lab":"Anthropic","cold":[[9,925,1324,3155],[10,1033,2054,4655],[11,1471,3298,5888],[12,1721,3515,5852],[13,1436,2455,4335],[14,2003,3273,5158],[15,2780,4313,7128],[16,3696,5560,8837],[17,5577,7553,11351],[18,8324,11735,17067],[19,15246,20428,27182]],"warm":[[10,1152,1954,3402],[11,1062,3006,5376],[12,1150,2903,4673],[13,1737,3305,5144],[14,1539,2816,4345],[15,1785,3071,4454],[16,2119,3305,4778],[17,2510,3735,5579],[18,2616,4319,6851],[19,3607,5956,9910]]},{"model":"gpt-6.1-sol","lab":"OpenAI","cold":[[8,992,1153,1465],[9,1014,1182,1559],[10,1089,1539,2779],[11,1180,1492,2354],[12,1354,1989,3804],[13,1536,2433,4147],[14,1634,2343,3990],[15,1546,2278,4049],[16,2587,4259,6262],[17,3562,5316,8133],[18,5702,7496,9552]],"warm":[[11,1190,1678,3859],[12,1296,1910,3754],[13,1448,1956,3235],[14,1602,2304,3747],[15,1997,3153,5456],[16,2307,3720,6645],[17,3043,4985,9149],[18,4023,6384,10885],[19,4980,7051,9902]]},{"model":"deepseek-v4-flash","lab":"DeepSeek","cold":[[8,336,486,859],[9,287,371,548],[10,301,389,585],[11,354,586,1041],[12,395,698,1325],[13,638,1044,1982],[14,1051,1545,3054],[15,1832,2382,3756],[16,3431,4901,8180],[17,5665,7500,10621],[18,13616,17857,26702],[19,55446,65576,72807]],"warm":[[8,172,191,274],[9,101,115,187],[10,341,442,688],[11,179,302,559],[12,241,410,783],[13,198,381,858],[14,354,754,1301],[15,432,877,1444],[16,518,922,1551],[17,679,1069,1790],[18,1008,1502,2416],[19,1303,1662,2296]]},{"model":"claude-sonnet-5-5","lab":"Anthropic","cold":[[8,553,772,991],[9,565,760,1400],[10,638,1039,3962],[11,585,978,3376],[12,841,2045,4406],[13,1099,2662,5091],[14,1234,2339,4565],[15,1573,2722,4997],[16,2310,3737,6648],[17,4034,5973,9776],[18,6927,10284,17100],[19,12196,19347,30775]],"warm":[[11,738,1078,2362],[12,540,723,1544],[13,1114,2530,4466],[14,1072,2136,3159],[15,1244,2217,3739],[16,1473,2536,4442],[17,1905,3005,5034],[18,2305,3567,6088],[19,3312,4691,7497]]}],"tools":[{"tool":"read","reach":100,"calls":40.2},{"tool":"write","reach":84.3,"calls":6.3},{"tool":"bash","reach":74.5,"calls":100},{"tool":"edit","reach":73.2,"calls":11.1},{"tool":"read_file","reach":63.4,"calls":13.4},{"tool":"grep","reach":60.8,"calls":18.8},{"tool":"glob","reach":57.5,"calls":4.9},{"tool":"web_search","reach":47.7,"calls":9.2},{"tool":"shell","reach":41.8,"calls":9.8},{"tool":"skill","reach":39.9,"calls":1.9},{"tool":"terminal","reach":36.6,"calls":62.5},{"tool":"execute_code","reach":31.4,"calls":11.6},{"tool":"write_file","reach":31.4,"calls":8.7},{"tool":"search_files","reach":28.8,"calls":3},{"tool":"webfetch","reach":28.8,"calls":0.7},{"tool":"skill_view","reach":27.5,"calls":4.9}],"sessions":[{"label":"1 request","sessions":96.399,"share":22.64,"duration":0},{"label":"2","sessions":1.359,"share":2.09,"duration":3},{"label":"3 to 5","sessions":1.261,"share":5.44,"duration":16},{"label":"6 to 11","sessions":0.566,"share":7.04,"duration":57},{"label":"12 to 22","sessions":0.227,"share":8.11,"duration":155},{"label":"23 to 45","sessions":0.112,"share":11.92,"duration":412},{"label":"46 to 90","sessions":0.052,"share":15,"duration":970},{"label":"91 to 181","sessions":0.019,"share":14.02,"duration":2340},{"label":"182 to 362","sessions":0.005,"share":9.45,"duration":4958},{"label":"363 to 724","sessions":0.001,"share":3.71,"duration":8153},{"label":"725 or more","sessions":0,"share":0.58,"duration":11949}],"pulse":{"hours":["2026-11-09T00","2026-11-09T01","2026-11-09T02","2026-11-09T03","2026-11-09T04","2026-11-09T05","2026-11-09T06","2026-11-09T07","2026-11-09T08","2026-11-09T09","2026-11-09T10","2026-11-09T11","2026-11-09T12","2026-11-09T13","2026-11-09T14","2026-11-09T15","2026-11-09T16","2026-11-09T17","2026-11-09T18","2026-11-09T19","2026-11-09T20","2026-11-09T21","2026-11-09T22","2026-11-09T23","2026-11-10T00","2026-11-10T01","2026-11-10T02","2026-11-10T03","2026-11-10T04","2026-11-10T05","2026-11-10T06","2026-11-10T07","2026-11-10T08","2026-11-10T09","2026-11-10T10","2026-11-10T11","2026-11-10T12","2026-11-10T13","2026-11-10T14","2026-11-10T15","2026-11-10T16","2026-11-10T17","2026-11-10T18","2026-11-10T19","2026-11-10T20","2026-11-10T21","2026-11-10T22","2026-11-10T23","2026-11-11T00","2026-11-11T01","2026-11-11T02","2026-11-11T03","2026-11-11T04","2026-11-11T05","2026-11-11T06","2026-11-11T07","2026-11-11T08","2026-11-11T09","2026-11-11T10","2026-11-11T11","2026-11-11T12","2026-11-11T13","2026-11-11T14","2026-11-11T15","2026-11-11T16","2026-11-11T17","2026-11-11T18","2026-11-11T19","2026-11-11T20","2026-11-11T21","2026-11-11T22","2026-11-11T23","2026-11-12T00","2026-11-12T01","2026-11-12T02","2026-11-12T03","2026-11-12T04","2026-11-12T05","2026-11-12T06","2026-11-12T07","2026-11-12T08","2026-11-12T09","2026-11-12T10","2026-11-12T11","2026-11-12T12","2026-11-12T13","2026-11-12T14","2026-11-12T15","2026-11-12T16","2026-11-12T17","2026-11-12T18","2026-11-12T19","2026-11-12T20","2026-11-12T21","2026-11-12T22","2026-11-12T23","2026-11-13T00","2026-11-13T01","2026-11-13T02","2026-11-13T03","2026-11-13T04","2026-11-13T05","2026-11-13T06","2026-11-13T07","2026-11-13T08","2026-11-13T09","2026-11-13T10","2026-11-13T11","2026-11-13T12","2026-11-13T13","2026-11-13T14","2026-11-13T15","2026-11-13T16","2026-11-13T17","2026-11-13T18","2026-11-13T19","2026-11-13T20","2026-11-13T21","2026-11-13T22","2026-11-13T23","2026-11-14T00","2026-11-14T01","2026-11-14T02","2026-11-14T03","2026-11-14T04","2026-11-14T05","2026-11-14T06","2026-11-14T07","2026-11-14T08","2026-11-14T09","2026-11-14T10","2026-11-14T11","2026-11-14T12","2026-11-14T13","2026-11-14T14","2026-11-14T15","2026-11-14T16","2026-11-14T17","2026-11-14T18","2026-11-14T19","2026-11-14T20","2026-11-14T21","2026-11-14T22","2026-11-14T23","2026-11-15T00","2026-11-15T01","2026-11-15T02","2026-11-15T03","2026-11-15T04","2026-11-15T05","2026-11-15T06","2026-11-15T07","2026-11-15T08","2026-11-15T09","2026-11-15T10","2026-11-15T11","2026-11-15T12","2026-11-15T13","2026-11-15T14","2026-11-15T15","2026-11-15T16","2026-11-15T17","2026-11-15T18","2026-11-15T19","2026-11-15T20","2026-11-15T21","2026-11-15T22","2026-11-15T23"],"eu":[8.92,5.65,4.6,8.22,8.42,12.66,17.61,34.33,51.52,57.54,65.2,60.41,62.17,59.43,59.79,59.98,50.02,39.45,37.06,30.77,30.24,25.21,31.23,27.94,12.29,5.97,5.84,6.31,12.07,12.4,18.64,33.71,56.04,71.67,82.4,75.16,75.82,77.84,76.59,76.43,60.06,53.12,33.39,25.71,31.01,31.94,29.98,18,12.37,8.46,6.63,7.91,9.48,14.97,19.63,35.63,50.67,62.25,69.9,57.88,66.66,78.62,71.89,71.63,62.46,49.24,39.23,35.89,29.81,28.47,26.59,19.26,21.3,13.54,13,15.44,17.57,18.77,26.07,37.38,59.27,69.14,80.59,76.28,75.96,80.28,84.84,73.96,59.28,45.39,33.72,32.4,34.37,35.9,27.95,27.6,17.74,14.69,12.39,9.03,9.88,12.07,21.41,40.27,60.78,76.03,79.09,76.54,69.38,70.86,71.24,67.69,43.79,36.49,33.1,28.05,28.76,28.51,28.85,22.43,16.72,10.42,9.56,7.47,8.78,10.98,13.5,19.94,22.42,30.45,29.76,27,20.8,21.9,21,24.24,22.97,18.95,20.42,24.18,23.83,23.57,21.66,16.88,13.5,8.32,7.42,5.42,5,6.09,5.83,8.8,12.96,20.17,21.72,21.32,26.02,26.01,27.38,24.98,22.74,26.26,22.87,25.79,19.91,25.7,20.47,16.85],"us":[1.81,3.69,4.2,3.72,1.96,1.42,1.96,2.69,2.18,2.27,1.2,1.67,2.17,2.07,3.2,2.64,3.52,2.27,3.03,2.45,2.44,2.57,3.66,3,3.12,3.11,3.19,3.11,2.37,3.06,3.89,3.23,2.12,2.01,1.82,2.19,2.69,1.95,2.56,3.76,4.13,5.81,2.54,2.97,2.44,2.41,2.76,2.69,1.94,2.67,3.57,2.86,3.02,3.33,1.72,2.43,2.05,2.4,1.74,1.8,4.8,6.21,6.7,6.32,3.71,6.33,3.87,2.24,4.15,3.89,4.19,3.04,2.84,2.1,2.67,2.68,2.36,3.18,2.94,3.42,1.97,2.77,2.15,2.94,2.78,1.84,3.86,2.75,3.32,2.83,3.07,2.17,1.72,2.76,3.24,2.43,2.36,2.12,2.61,2.15,1.22,2.4,2.16,2.99,3.42,2.05,2.57,2.16,2.09,2.75,3,3.67,3.28,2.96,2.42,2.59,2.35,2.19,2.02,3.22,2.9,3.84,2.67,1.59,1.43,1.91,1.7,1.87,1.94,0.67,0.61,0.82,1.26,2.44,2.56,2.47,1.8,2.06,2.35,2.57,2.44,2.07,1.67,2.12,3.33,3.42,3.22,3.52,1.57,2.81,1.42,0.67,0.89,1.15,0.69,1.17,1.48,3.25,21.12,6.75,1.92,3.13,3.16,3.45,2.29,2.78,2.54,2.09],"ap":[3.9,5.28,5.67,8.46,7.41,7.81,11.15,11.74,14.72,13.13,21.04,22.4,21.58,9.09,7.68,5.81,5.27,3.23,2.64,2.36,2.52,1.78,1.77,2.17,4.28,6.15,8.29,8.59,10.3,8.51,12.39,12.35,9.19,12.51,8.33,8.1,6.58,5.76,4.1,4.56,3.55,3.36,1.91,1.63,2.8,2.13,1.29,2.18,2.85,4.52,8.07,12.94,12.13,12.77,12.15,10.62,10.61,11.35,8.54,8.41,9.26,8.33,5.56,3.84,4.74,2.49,1.84,1.3,1.32,1.12,1.2,1.52,2.26,5.03,7.54,12.01,9.38,8.84,12.36,11.41,11.61,7.59,7.92,7.2,7.1,10.93,11.3,11.36,10.94,9.31,7.81,7.09,7.24,6.82,6.09,6.68,8.43,9.31,11.45,11.52,8.2,6.51,8.54,7.4,11.02,11.13,9.4,6.64,6.3,5.01,6,4.92,2.86,3.63,2.5,1.49,1.21,1.67,2.05,1,3.36,1.46,1.94,2.3,2.25,3.65,4.19,4.61,3.25,4.71,4.01,3.39,4.45,3.5,4.65,5.24,3.35,3.77,2.11,2.22,2.03,2.1,2.36,2.38,1.86,2.44,3.26,3.8,3.24,2.48,2.38,4.47,4.11,2.74,2.36,2.65,3.88,3.29,3.57,2.96,3.17,3.3,2.76,2.22,0.83,0.98,0.9,1.62],"err":[3.734,0.46,0.211,3.186,0.145,0.142,0.095,0.122,0.083,0.076,0.149,0.142,0.353,0.364,0.663,0.787,0.795,0.983,0.823,0.953,0.843,1.278,0.963,0.703,0.779,0.918,0.877,2.021,0.804,0.614,0.595,0.479,0.36,0.34,5.639,0.546,0.481,1.162,1.267,1.13,1.416,0.941,1.923,1.994,1.249,1.134,1.061,1.305,0.915,1.027,0.749,1.626,0.698,0.729,0.833,0.635,0.756,0.632,0.76,1.109,1.31,1.48,1.755,2.314,2.47,3.182,3.117,0.977,1.27,0.603,0.511,0.376,0.182,0.108,0.055,0.78,0.253,0.07,0.148,0.051,0.237,0.348,0.486,0.293,0.12,0.51,0.878,1.1,1.305,0.462,0.828,0.382,0.687,0.711,0.404,0.507,0.206,0.084,0.08,0.886,0.128,0.046,0.088,0.093,0.136,0.133,0.13,0.104,0.246,0.716,0.256,0.306,0.922,0.779,0.734,0.447,0.528,0.537,0.544,0.636,0.073,0.054,0.044,1.104,0.066,0.124,0.025,0.036,0.033,0.047,0.064,0.083,0.173,0.271,0.222,1.184,13.181,0.616,0.484,0.504,1.022,0.287,0.312,0.247,0.201,0.128,0.041,1.206,0.139,0.032,0.049,0.048,0.148,0.064,0.017,0.035,0.417,0.373,0.484,0.582,0.875,0.9,0.524,0.454,0.728,0.279,0.467,0.078],"ttft":[1266,1332,1240,1437,1562,1445,1405,1668,1362,1413,1529,1255,1351,1610,1597,1642,1557,1719,1868,1841,1905,1661,1952,1346,1166,1291,1227,1209,1398,1600,1687,1771,1584,1616,1669,1578,1502,1674,1765,1599,1689,1537,1382,1366,1348,1282,1209,1100,1086,1223,1425,1198,1397,1201,1319,1522,1298,1278,1293,1407,1320,1172,1537,1589,1467,1376,1590,2452,1557,1159,1379,1508,1120,1221,1068,719,1395,1352,1323,1472,1315,1406,1376,1223,1264,1345,1687,1654,1521,1555,1395,1310,1250,1323,1329,1119,1074,941,1006,922,1317,1542,1463,1420,1363,1408,1490,1355,1136,1420,1602,2004,1645,1577,1460,1239,1262,1201,1056,993,977,1263,1148,1198,1245,1502,1674,1326,1080,1254,1027,925,1167,1100,1108,983,1042,1288,1248,1202,1287,1252,1075,1016,1211,1228,1113,1447,1220,1160,611,1369,1329,1343,1428,1412,1500,1891,2000,1692,1532,1499,1407,1607,1400,1473,1417,1289]},"heat":[{"host":"azure","label":"Azure","v":[0,0,0,0,0,0,0.01,0.03,0.02,0.03,0.03,0.03,0.06,0.1,0.03,0.07,0.02,0.01,0.04,0.01,0.03,0.09,0.05,0.05,0,0,0,0.05,0.01,0,0.08,0.05,0.03,0.04,0.02,0.05,0.01,0.03,0.18,0.01,0.03,0.18,0.02,0,0.01,0.01,0,0.04,0,0.02,0.03,0.03,0.04,0.01,0.01,0.01,0.03,0.01,0.13,0.01,0.01,0.1,0.71,0.95,0.18,0.06,0,0.03,0,0,0,0,0,0.04,0.04,0,0,0,0.03,0.04,0.02,0.01,0.03,0.04,0.01,0.03,0.01,0.01,0.03,0,0.02,0.01,0.04,0.07,0.01,0.22,0.06,0,0.05,0.02,0.01,0.01,0.01,0,0.01,0,0.02,0.02,0.01,0.03,0.05,0.13,0.09,0.01,0.01,0.23,0.03,0,0,0,0,0.03,0,0.02,0,0,0.04,0.01,0,0,0.01,0,0,0,0.02,0.01,0.05,0.06,0.03,0,0,0,0,0,0,0,0,0,0,0.01,0,0.03,0,0,0,0.02,0,0,0,0.02,0.01,0.01,0,0,0,0,0.01,0],"err":0.05,"prevErr":3.17},{"host":"bedrock","label":"Bedrock","v":[0,0.04,0,0.03,0.04,0,0,0.01,0,0.02,0.01,0,0,0,0,0.02,0,0.01,0,0,0,0.01,0,0,0,0,0,0,0.02,0,0,0.03,0.01,0,29.45,0.02,0,0,0,0,0.01,0,0,0,0,0,0,0,0,0.03,0,0,0.07,0,0,0.03,0.02,0,0.1,0.01,0,0.01,0,0.01,0,0,0,0.01,0,0.01,0.01,0,0,0,0.01,0,0,0.01,0,0.01,0.01,0,0,0.1,0.01,0,0,0,0,0,0,0.01,0,0,0,0.19,0.01,0.06,0.08,0,0.01,0.01,0,0.01,0,0,0,0,0,0,0,0.04,0.95,0.72,0.04,0.32,0.44,0.63,0.55,0.09,0,0,0.03,0,0,0.02,0,0,0,0,0.03,0.01,0.01,0.04,0.03,6.02,48.85,0.03,0,0,0,0.01,0,0.03,1.71,1.94,0.03,0.06,0,0.04,0,0.05,0.03,0,0.02,0,0,0.19,0.02,0,0,0.02,0,0.04,0,0,0,0.07],"err":2.3,"prevErr":0.24},{"host":"sference","label":"Sference","v":[23.23,1.45,0,0,0.29,0,0,0.01,0.08,0.06,0.07,0.08,0.19,0.05,0.27,0.22,0.06,0.03,0.01,0,0,0,0.17,0,0,0,0,0,0,0,0.02,0.03,0.12,0.1,0.31,0.12,0.25,0.45,0.2,0.04,0.04,0.01,0,0,0.01,0.01,0.03,0,0,0,0.03,0,0,0,0,0.01,0.03,0.11,0.04,0.08,0.13,0.06,0.02,0.03,0.05,0.01,0.02,0.05,0.36,0,0,0,0,0,0,0,0,0.06,0.01,0.09,2,3.36,4.82,3.25,0.05,3.87,6.94,9.3,10.83,0.04,0,0.01,0,0.02,0,0,0.04,0,0,0,0,0,0,0.04,0.04,0.17,0.01,0.01,0.12,0.16,0.11,0.1,0.08,0.02,0,0,0.04,0.13,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.03,0,0,0.02,0,0,0.01,0,0,0.24,1.1,0,0,0,0,0,0,0.07,0,0,0,0,0,0,0,0,0,0,0.02,0.14,0.02,8.63,0,0,0],"err":0.68,"prevErr":3.09},{"host":"melious","label":"Melious","v":[34.56,15.68,0,0,0,0,0,0,0,0,0.11,0,0.03,0.45,3.84,0.61,0.03,0.59,0,0,0,0,0.71,0,0,0,0,0,0,0,0,0,0.02,0.07,0.18,0.04,0.06,0.56,0.05,0.03,0.03,0,0,0,0,0,0,0,0,0,0,0.08,0.09,0,0,0.02,0,0,0,0,0,0.01,0.01,0.01,0.01,0,0,0,0,0,0,0.69,0,0,0,0,0,0,0,0,0,0,0.03,0,0.01,0.01,0.04,0,0,0.04,0.12,0,0,0,0,0,0,0,0,0,0,0,0,0.02,0,0.05,0.01,0.01,0.54,0.42,0.02,0.06,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0.32,0,0,0,0,0.07,0,0,0,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.04,0.53,11.85,0,0.05,0,0,0,0,0,0,0,0,0],"err":0.26,"prevErr":9.94},{"host":"deepinfra","label":"DeepInfra","v":[0.03,0,0,0,0,0.04,0.05,0,0.01,0.02,0.03,0.02,0,0.03,0.18,0,0.09,0.11,0.46,0,2.42,0.7,0.74,1.71,0.2,0,0,0,0,0,0.14,0.07,0.01,0.01,0.03,0.06,0.03,0,0,0,0.1,0.5,0.02,0,0,0,0,0.29,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.1,0.02,0,0.04,0,0,0,0.04,0.17,0,0.04,0,0.03,0.02,0.01,0.01,0.01,0.2,0.02,0.01,0,0,0.01,0.02,0.01,0,0.01,0.01,0.01,0,0.01,0.01,0.01,0,0.94,0.01,0.02,0,0.01,0,0.03,0.09,0,0.02,0,0.01,0.02,0.04,0.2,0.01,0.01,0.06,0.05,4.3,0.02,0.05,0,0,0.12,0,0,0.01,0,0.02,0,0.02,0,0.05,0.11,0.04,0.02,0.79,0,0,0,0,0,0.16,0,0,0,0,1.02,0,0,0,0.03,0,0,0,0,0,0,0.13,0.13,0.93,0,0.19,0,0.05,0.36,0.04],"err":0.08,"prevErr":0.32},{"host":"vertex_claude","label":"Vertex (Claude)","v":[0,0,0,0,0,0,0,0.04,0.07,0.02,0,0,0.12,0,0.12,0.26,0.02,0,0,0.05,0.93,0.03,0,0,0,0,0,0,0,0,0,0,0,0,0,0.02,0.15,0,0.17,0.04,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.03,0,0,0,0,0,0,0.02,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.01,0,0,0.01,0.02,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.49,0.03,0,0,0,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"err":0.02,"prevErr":0.05},{"host":"anthropic","label":"Anthropic","v":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.02,0,0,0,0.03,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.03,0,0,0,0,0,0,0.08,0,0,0,0,0,1.12,5.08,0,2.05,7.57,1.08,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"err":0.11,"prevErr":0},{"host":"vertex_gemini","label":"Vertex (Gemini)","v":[0.01,0,0,0,0.01,0,0.14,0.01,0.02,0.05,0,0,0.04,0.01,0.01,0.02,0.02,0.03,0.02,0.02,0.03,0.02,0.02,0.04,0.07,0.03,0.03,0.02,0,0.02,0.03,0,0.02,0.31,1.81,0.64,0.06,0.57,0.41,0.09,0.12,0.06,0.11,0.08,0,0,0,0.02,0,0,0,0.01,0.03,0.01,0.04,0.02,0.05,0.01,0.06,0.04,0.05,0.09,0.02,0.04,0.02,0.09,0.04,0.03,0.03,0.08,0.04,0.22,0.02,0,0.03,0.02,0.02,0.04,0.02,0.02,0.06,0.02,0.02,0.07,0.09,0.04,0.01,0.19,0.08,0.05,0.07,0.01,0.03,0.02,0.03,0.01,0.02,0.01,0,0.52,0.13,0.01,0.03,0.01,0.08,0.03,0.03,0.06,0.01,0.02,0.02,0.01,0.03,0.05,0.05,0.07,0.01,0.01,0,0.01,0,0,0,0.05,0.02,0.09,0.01,0.01,0.03,0.04,0,0.02,0.02,0.03,0.02,0,0,0.01,0.04,0,0,0.02,0,0,0,0,0,0,0,0,0,0.01,0,0.01,0.01,0,0,0,0.02,0,0,0,0,0,0,0,0,0],"err":0.07,"prevErr":0.06},{"host":"openai","label":"OpenAI","v":[0.05,0.02,0,0.12,0.01,0.02,0.11,0.03,0,0,0,0.03,0.06,0.01,0.24,0.47,0.39,0.31,0.14,0.24,0.06,0.04,0.03,0.34,0.53,0.33,1.69,0.29,0.27,0.04,0.03,0.03,0.05,0.02,0.05,0.02,0.07,0.02,0.52,0.21,0.1,0.02,0.03,0.29,0.47,0.89,1.1,0.54,0.02,0.01,0,0.12,0.28,0.04,1.08,0.02,0.16,0.02,0,0.01,0.08,0.22,0.09,0.12,0.36,0.5,1.35,0.29,0.42,0.66,0.13,0.41,0.35,0.2,0.1,0.15,0.09,0.08,0.11,0,0.07,0.03,0.19,0.12,0.2,0.2,0.26,0.16,0.51,0.8,0.91,0.39,0.39,0.19,0.43,0.5,0.1,0.1,0.06,0.08,0.03,0.1,0.03,0.19,0.01,0.04,0.1,0.05,0.04,0.01,0.02,0.01,0.04,0.04,0.1,0.06,0.1,0.06,0.05,0.02,0.03,0.02,0.03,0.03,0,0.03,0.01,0.06,0.01,0,0.02,0,0,0,0.01,0,0.01,0.03,0.01,0.03,0.02,0,0,0,0,0,0,0.17,0,0,0.01,0.03,0,0,0,0.01,0,0,0.01,0.06,0,0.01,0,0.01,0.03,0.01,0.1,0.01],"err":0.16,"prevErr":0.23},{"host":"tensorx","label":"TensorX","v":[0.13,0.28,0,0.36,0.36,0,0.39,0.15,0.73,0.58,0.09,1.09,0.07,1.79,0.93,0.41,0.19,0.4,0.23,0.61,0.25,0.15,0.07,0.03,0.05,0,0.19,0.25,0.77,0.18,0.23,0.24,0.59,0.94,0.23,0.48,0.13,0.09,0.06,0.94,0.53,0.02,0.65,0.25,0,0.04,0.25,0.34,0.03,3.78,0.67,0.25,0.05,0.62,0.42,0.7,0.28,0.02,0.56,0.24,0.43,0.09,0.28,0.32,0.03,0.3,3.82,0.01,15.48,0.12,0.15,0.96,0.3,0.08,0.03,0.45,0.07,0.65,0,0.07,0.27,0.88,0.24,0.07,0.02,0.09,0.6,0.25,0.3,0.83,0.11,0.45,0,1.5,0,0.62,0.36,0.73,3.25,0.37,0.62,0,0.2,0.1,1.41,0.14,0.01,0.01,0.09,0.36,0.32,0.97,0.36,0.36,0.5,0.04,0.35,0.06,1.06,0.29,0.22,0.29,0.58,0,1.25,0,0.04,0.02,0.09,0.33,0.33,0.72,0.03,0,0.03,0.08,0.57,0.51,0.03,0.06,0.46,0.02,0.66,0.05,0.03,0.5,0.16,0.05,0,0,0,0,0,0.58,0,0,0,0.28,0,0,0.04,0.55,0.07,0.05,2.75,0.15,0,0],"err":0.49,"prevErr":0.62},{"host":"deepseek","label":"DeepSeek","v":[0,0,0,0,0,0.03,0,0.03,0,0.02,0,0,0.02,0,0,0.01,0,0,0,0,0,0,0,0,0,0.13,0,0,0,0,0.03,0,0.17,0.57,0.27,0.04,0,0,0,0.03,0,0,0,0.02,0,0.04,0,0.02,0,0.13,0,0.08,0,0,0.14,0.05,0,0.34,0,0,0,0,0,0,0.03,0,0,0.06,0.02,0,0,0.04,0,0,0,0.05,0,0.04,0,0,0,0.08,0,0,0,0,0,0,0,0,0,0.18,0,0,0,0,0,0,0.11,0.1,0.05,0,0,0,0,0.03,0,0.02,0,0,0,0,0,0,0,0,0,0,0,0,0,0.04,0.03,0.18,0,0,0,0,0,0.05,0.03,0,0,0,0,0,0.04,0,0,0.03,0,0,0,0,0,0.11,0.19,0.19,0.1,0.1,0.07,0,0,0.06,0,0.18,0,0,0,0,0,0.18,0.15,0.16,0,0,0,0],"err":0.03,"prevErr":0.79},{"host":"nvidia","label":"NVIDIA","v":[0.16,0.06,0.06,0,0,0.05,0.12,0.17,0.19,0.19,0.05,0,0.19,0,0.07,0,0.82,0.27,0,13.36,2.79,0.09,0.31,0.1,0,0.06,0,0.23,0,0.04,0.2,0.26,0.12,0.22,0.19,0.15,0.15,0,0.31,0.15,0.34,0.55,0.19,0.21,0.2,0.75,0.42,0.1,0.07,0.29,0,0.09,2.04,0.12,0.37,0,0.08,3.16,18.06,23.05,3.87,2.63,2.61,3.53,3.02,2.05,1.9,2.04,3.38,2.45,1.54,1.15,1.75,0.54,0,0.08,2.24,0.17,0.05,0,2.69,2.3,0,0,0,0.06,0.48,0.19,0.06,0,0,0.31,0.06,0.06,0.36,0.15,0.1,0.19,0.4,0.23,0.17,0.07,0.11,0.11,0,0.15,0.06,0.2,0.17,0.43,0.47,0.68,1.02,0.67,0.38,0.22,0.23,6.3,9.28,1.12,0.34,0.06,0.48,0.26,0.16,0.06,0.1,0.06,0.11,0,0,0.09,0.23,0.22,0.18,0.28,0.44,0.2,0.18,5.47,8.99,0.06,0.36,0.3,0.93,0.17,0.2,0,0.15,0.11,0.28,0.26,0.05,0.12,0.1,0.08,0.84,0.72,0.22,2.14,17.67,4.44,0.35,0.8,1.24,0.57,1.66,0.78],"err":1.03,"prevErr":1.48},{"host":"fireworks","label":"Fireworks","v":[0,0,0.72,0,0,0,0,0,0,0,0,0,0,0,0,0.56,0,0,0,0,0,0,0,0,0,0,0.23,0.15,0,0,0,0,0,0,0,0,0,12.38,0,0,0,0,0,0.17,0,0,0,0.12,0,0,0,0,0,0,0,0,0,0,0,0,0,5.46,0,0,0,0,0.46,0,0,0,0,0,0,0,0,0,0.11,0,0,0,0.37,0,0,0,0,0.02,0,0,0,0,0,0,0.06,0,0,0,7.06,11.59,0,0,0,0,2.25,0,0,0,0,0,0,0,0,0,1.61,0,0,0,1.48,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"err":0.33,"prevErr":2.48},{"host":"mistral","label":"Mistral","v":[0,0,0,0,0,0.12,0.18,0,0.18,0,0,0,0,0,0,2.85,0.61,0,0.1,0.13,0,0.1,0,0,0,0.55,0,0,0,0.17,0.13,0,0,0,0,1.03,0,0,0.17,0.59,1.08,0,0,0,0,0,0.18,0.07,0,0.16,0,0,0.13,0,0.14,0.14,0,0.42,2.4,0.56,3.65,2.53,4.66,0.97,0.73,0,0,0.19,0,0.05,0,0,0.09,0.12,0,0,0,0.23,0,0,0.35,1.64,0.48,2.03,1.7,2.72,5.18,4.66,1.34,0.46,0.25,0,0.1,0,0,0,0.65,0,0,0,0,0,0,0,0.05,0.75,0.43,0,0,0.92,0.25,0.29,0.58,3.68,0,0,0,0,0.06,0.07,0,0,0,0,0,0,0,0,0.1,0,0,0,0,0,0,0.08,0,0,0,0,0,0.07,0.08,0,0,0,0,0,0,0,0,0,0.11,0,0,0.11,0,0,0,0,3.05,0.57,3.41,0,0.05,0.09,0,0],"err":0.69,"prevErr":0.11},{"host":"novita","label":"Novita","v":[0,0,0,0,0,2.29,0,0.03,0.04,0.01,0.07,0,0,0,0.46,0,0,0,0,0,0,0.64,0,0,0,0,0,0,0,7.46,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1.66,1.64,0,0,0,0,0,0,0,0,4.61,0,0,0,0,0,0,0,0,0,0,0,0.51,0,0,0,0,1.04,0,0,0,0,0,0,0,2.48,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1.5,0,0,0,0,0,0,0,0.09,0,0,0,0,0,0,0,0,0,0,0,0.18,0,0,0,0.67,0,0,0,0,0,0,0,0.12,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1.01,0,0,0,0,0,0,0,0.15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"err":0.04,"prevErr":2.5},{"host":"zai","label":"Z.ai","v":[0,0,0,null,null,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,null,null,null,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,null,null,0,0,0,0,0,0,0,0.29,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,12.55,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,null,null,0,0,0,0,0,0,0,0,0,0,0,0,0,0,null,0,0,0,0,0,0,0,0,0,0,0,null,null,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"err":0.11,"prevErr":0}]},"charts":[{"id":"token-flow","parent":"top-model","chart":"flow","topic":"Rankings","title":"Where the week's tokens went","lede":"Apps on the left, labs in the middle, hosts on the right. Follow a lab to see who sends to it and who runs it. One lane is 1% of the week's tokens.","short":"DeepSeek took a third of all tokens.","claim":"DeepSeek models took 34.1% of tokens on the Requesty gateway in week 46 (Nov 9 to 15, 2026); Azure served the most of any host, 15.3%, and the OpenAI Python SDK sent the most of any app, 21.3%.","soWhat":"Every token leaves an app, goes to a lab's model and is served by a host, and the three rarely line up one to one. DeepSeek's tokens alone ran on DeepInfra, Sference, Melious and DeepSeek, plus smaller hosts.","action":{"text":"Route one model across several hosts","href":"https://docs.requesty.ai/features/load-balancing"},"metric":{"value":34.1,"unit":"pct","denominator":"of tokens, each team capped at 2%"},"footnote":"Share of tokens by lab, each team capped at 2% · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"pct","rows":[{"label":"DeepSeek","value":34.07,"highlight":true},{"label":"Anthropic","value":24.4,"highlight":false},{"label":"OpenAI","value":19.69,"highlight":false},{"label":"Z.ai","value":6.62,"highlight":false},{"label":"Google","value":4.47,"highlight":false},{"label":"NVIDIA","value":3.59,"highlight":false},{"label":"MiniMax","value":0.94,"highlight":false}]},"posts":{"x":["Where last week's tokens went on Requesty (Nov 9 to 15): DeepSeek models took 34.1%, Anthropic 24.4%. Azure served the most of any host (15.3%), and the OpenAI Python SDK sent the most of any app.","DeepSeek took 34.1% of all tokens on Requesty last week, served by DeepInfra, Sference, Melious and more. One lab, many places to run it (Nov 9 to 15).","Apps, labs, hosts: the busiest path on Requesty last week ran from the OpenAI Python SDK (21.3% of tokens) to DeepSeek (34.1%) to Azure (15.3%). Nov 9 to 15."],"linkedin":"DeepSeek models took 34.1% of tokens on the Requesty gateway in week 46 (Nov 9 to 15, 2026); Azure served the most of any host, 15.3%, and the OpenAI Python SDK sent the most of any app, 21.3%.\n\nEvery token leaves an app, goes to a lab's model and is served by a host, and the three rarely line up one to one. DeepSeek's tokens alone ran on DeepInfra, Sference, Melious and DeepSeek, plus smaller hosts.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"agent-tools","parent":"coding-agents","chart":"tools","topic":"Apps","title":"What agents do all day","lede":"The most common tool names in model tool calls. Bar: how many teams use the tool. Dot: how often it is called.","short":"Agents call bash 2.5× as often as read.","claim":"read was the tool the most teams' models called on the Requesty gateway in week 46 (Nov 9 to 15, 2026); bash was called 2.5x as often.","soWhat":"Agents spend their turns reading files, running commands and editing code. Tool results go straight back into the next prompt, which is why agent prompts are long and mostly cached.","action":{"text":"Run Claude Code through Requesty","href":"https://docs.requesty.ai/integrations/claude-code"},"metric":{"value":2.5,"unit":"x","denominator":"bash calls for every read call"},"footnote":"Teams calling each tool, top tool = 100 · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"count","rows":[{"label":"read","value":100,"highlight":true},{"label":"write","value":84.3,"highlight":false},{"label":"bash","value":74.5,"highlight":true},{"label":"edit","value":73.2,"highlight":false},{"label":"read_file","value":63.4,"highlight":false},{"label":"grep","value":60.8,"highlight":false},{"label":"glob","value":57.5,"highlight":false},{"label":"web_search","value":47.7,"highlight":false},{"label":"shell","value":41.8,"highlight":false},{"label":"skill","value":39.9,"highlight":false}]},"posts":{"x":["What agents did all day on Requesty last week (Nov 9 to 15): read, write, bash, edit, read_file. read is the tool the most teams call; bash is called most often, 2.5x as much.","More teams' agents call read than any other tool on Requesty. But bash wins on volume: 2.5x as many calls last week (Nov 9 to 15). Agents read, then they run things.","Top tool names in model tool calls on Requesty last week, by teams using them: read, write, bash, edit, read_file, grep. Nov 9 to 15."],"linkedin":"read was the tool the most teams' models called on the Requesty gateway in week 46 (Nov 9 to 15, 2026); bash was called 2.5x as often.\n\nAgents spend their turns reading files, running commands and editing code. Tool results go straight back into the next prompt, which is why agent prompts are long and mostly cached.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"session-length","parent":"coding-agents","chart":"sessions","topic":"Apps","title":"A few long sessions burn most of the tokens","lede":"Sessions grouped by how many requests they contain. Grey: share of sessions. Blue: share of tokens.","short":"0.2% of sessions used 55% of tokens.","claim":"Sessions of 23 requests or more were 0.19% of sessions on the Requesty gateway in week 46 (Nov 9 to 15, 2026), but used 54.7% of tokens.","soWhat":"96.4% of sessions are a single request and use 22.6% of tokens. The bill is set by the few long agent loops, so that is where caching and model choice pay off.","action":{"text":"Turn on auto caching","href":"https://docs.requesty.ai/features/auto-caching"},"metric":{"value":54.7,"unit":"pct","denominator":"of tokens, in sessions of 23+ requests"},"footnote":"Share of tokens by requests per session · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"pct","rows":[{"label":"1 request","value":22.64,"highlight":false},{"label":"2 requests","value":2.09,"highlight":false},{"label":"3 to 5 requests","value":5.44,"highlight":false},{"label":"6 to 11 requests","value":7.04,"highlight":false},{"label":"12 to 22 requests","value":8.11,"highlight":false},{"label":"23 to 45 requests","value":11.92,"highlight":true},{"label":"46 to 90 requests","value":15,"highlight":true},{"label":"91 to 181 requests","value":14.02,"highlight":true},{"label":"182 to 362 requests","value":9.45,"highlight":true},{"label":"363 or more requests","value":4.29,"highlight":true}]},"posts":{"x":["96.4% of sessions on Requesty last week were a single request. Sessions of 23+ requests were 0.19% of sessions but used 54.7% of all tokens (Nov 9 to 15).","A few long sessions burn most of the tokens: on Requesty last week, sessions of 23 requests or more were 0.19% of sessions and 54.7% of tokens. Agent loops set the bill.","Single-request sessions: 96.4% of sessions, 22.6% of tokens. Sessions of 23+ requests: 0.19% of sessions, 54.7% of tokens. Requesty, Nov 9 to 15."],"linkedin":"Sessions of 23 requests or more were 0.19% of sessions on the Requesty gateway in week 46 (Nov 9 to 15, 2026), but used 54.7% of tokens.\n\n96.4% of sessions are a single request and use 22.6% of tokens. The bill is set by the few long agent loops, so that is where caching and model choice pay off.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"latency-spread","parent":"speed-up","chart":"latency","topic":"Speed","title":"Latency is a distribution","lede":"Each ridge is one model's week on a log scale, with the median and the point where the slowest ten percent start.","short":"glm-5.3-flash was quickest to start.","claim":"glm-5.3-flash had the quickest median first token of the 12 busiest models on the Requesty gateway in week 46 (Nov 9 to 15, 2026), 746 ms; for deepseek-v4-flash the slowest 1% of requests waited 36x its median.","soWhat":"A median hides the wait your slowest users feel. Look at p90 and p99 before you pick a model for anything interactive, and set timeouts from the tail, not the middle.","action":{"text":"Let the gateway pick the fastest host","href":"https://docs.requesty.ai/features/latency-routing"},"metric":{"value":746,"unit":"ms","denominator":"median time to first token"},"footnote":"Median time to first token, streaming requests · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"ms","rows":[{"label":"glm-5.3-flash","value":746,"highlight":true},{"label":"deepseek-v4.1-flash","value":834,"highlight":false},{"label":"deepseek-v4-flash","value":1140,"highlight":false},{"label":"gpt-6-luna","value":1677,"highlight":false},{"label":"claude-haiku-5-5","value":1755,"highlight":false},{"label":"nemotron-3-ultra-550b-a55b","value":1868,"highlight":false},{"label":"claude-sonnet-5","value":2323,"highlight":false},{"label":"gpt-5.6-luna","value":2463,"highlight":false},{"label":"gpt-6.1-terra","value":2711,"highlight":false},{"label":"claude-sonnet-5-5","value":2727,"highlight":false}]},"posts":{"x":["Quickest median first token among the 12 busiest models on Requesty last week: glm-5.3-flash 746 ms, deepseek-v4.1-flash 834 ms, deepseek-v4-flash 1.14 s (Nov 9 to 15).","Latency is a distribution. deepseek-v4-flash had a median first token of 1.14 s on Requesty last week, but its slowest 1% waited 41 s, 36x longer (Nov 9 to 15).","glm-5.3-flash started fastest of the busiest models on Requesty last week: 746 ms to the first token at the median, 3.3 s at p90. Measured at the gateway, Nov 9 to 15."],"linkedin":"glm-5.3-flash had the quickest median first token of the 12 busiest models on the Requesty gateway in week 46 (Nov 9 to 15, 2026), 746 ms; for deepseek-v4-flash the slowest 1% of requests waited 36x its median.\n\nA median hides the wait your slowest users feel. Look at p90 and p99 before you pick a model for anything interactive, and set timeouts from the tail, not the middle.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"same-model-hosts","parent":"speed-up","chart":"race","topic":"Speed","title":"Same weights, different metal","lede":"One model, many hosts, very different stopwatch. Pick a model and see every host that served it last week.","short":"Same model, 4.6× apart by host.","claim":"The same model, deepseek-v4.1-flash, reached its first token 4.6x sooner on Sference (375 ms median) than on TensorX (1.71 s) on the Requesty gateway in week 46 (Nov 9 to 15, 2026).","soWhat":"Same weights, different metal. The host you land on can matter more than the model you pick, and the busiest host for deepseek-v4.1-flash, Melious, was not the fastest.","action":{"text":"Let the gateway pick the fastest host","href":"https://docs.requesty.ai/features/latency-routing"},"metric":{"value":4.6,"unit":"x","denominator":"slowest vs fastest host, median time to first token, deepseek-v4.1-flash"},"footnote":"deepseek-v4.1-flash, median time to first token by host · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"ms","rows":[{"label":"Sference","value":375,"highlight":true},{"label":"Fireworks","value":565,"highlight":false},{"label":"Melious","value":670,"highlight":false},{"label":"Together","value":688,"highlight":false},{"label":"DeepSeek","value":1115,"highlight":false},{"label":"Runware","value":1352,"highlight":false},{"label":"DeepInfra","value":1633,"highlight":false},{"label":"TensorX","value":1714,"highlight":false}]},"posts":{"x":["Same weights, different metal. deepseek-v4.1-flash on Requesty last week: 375 ms to the first token on Sference, 1.71 s on TensorX. 4.6x apart for the same model (Nov 9 to 15).","8 hosts served deepseek-v4.1-flash on Requesty last week. The fastest, Sference, started in 375 ms at the median; the slowest took 1.71 s. Pick the host, not just the model.","The busiest host for deepseek-v4.1-flash on Requesty last week was Melious (31% of its tokens). Fastest to the first token: Sference, 375 ms. Nov 9 to 15."],"linkedin":"The same model, deepseek-v4.1-flash, reached its first token 4.6x sooner on Sference (375 ms median) than on TensorX (1.71 s) on the Requesty gateway in week 46 (Nov 9 to 15, 2026).\n\nSame weights, different metal. The host you land on can matter more than the model you pick, and the busiest host for deepseek-v4.1-flash, Melious, was not the fastest.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"host-form","parent":"rescues","chart":"hostForm","topic":"Reliability","title":"Host form","lede":"A form guide for the companies that run the models: provider errors per host for twelve weeks, then this week against last. A clean week shows a green dot.","short":"9 of 15 hosts had a clean week.","claim":"9 of the 15 busiest hosts on the Requesty gateway kept provider errors under 0.3% in week 46 (Nov 9 to 15, 2026); Bedrock had the most, 2.30%.","soWhat":"A host's form changes week to week, and last week's clean record says little about next week. Spread traffic across hosts so one bad week stays someone else's problem.","action":{"text":"Set up a fallback policy","href":"https://docs.requesty.ai/features/fallback-policies"},"metric":{"value":9,"unit":"count","denominator":"of the 15 busiest hosts kept provider errors under 0.3%"},"footnote":"Provider error rate by host, outline: the week before · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"pct","rows":[{"label":"Bedrock","value":2.3,"prev":0.24,"highlight":true},{"label":"NVIDIA","value":1.03,"prev":1.48,"highlight":false},{"label":"Mistral","value":0.69,"prev":0.11,"highlight":false},{"label":"Sference","value":0.68,"prev":3.09,"highlight":false},{"label":"TensorX","value":0.49,"prev":0.62,"highlight":false},{"label":"Fireworks","value":0.33,"prev":2.48,"highlight":false},{"label":"Melious","value":0.26,"prev":9.94,"highlight":false},{"label":"OpenAI","value":0.16,"prev":0.23,"highlight":false},{"label":"Anthropic","value":0.11,"prev":0,"highlight":false},{"label":"DeepInfra","value":0.08,"prev":0.32,"highlight":false}]},"posts":{"x":["Host form on Requesty last week (Nov 9 to 15): 9 of the 15 busiest hosts kept provider errors under 0.3%. Most errors: Bedrock, 2.30%, from 0.24% the week before.","Which hosts had a clean week on Requesty? Azure, Melious, DeepInfra, Vertex (Claude), Anthropic and 4 more stayed under 0.3% provider errors (Nov 9 to 15).","Provider error rate by host, Requesty, Nov 9 to 15: Bedrock 2.30%, NVIDIA 1.03%, Mistral 0.69%. The other 12 were lower."],"linkedin":"9 of the 15 busiest hosts on the Requesty gateway kept provider errors under 0.3% in week 46 (Nov 9 to 15, 2026); Bedrock had the most, 2.30%.\n\nA host's form changes week to week, and last week's clean record says little about next week. Spread traffic across hosts so one bad week stays someone else's problem.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"host-weather","parent":"rescues","chart":"heat","topic":"Reliability","title":"The week, hour by hour","lede":"Each cell is one hour on one host: the share of requests that came back as a rate limit, overload, timeout or server error.","short":"Bedrock's worst hour: 48.9% errors.","claim":"The worst hour for any busy host on the Requesty gateway in week 46 (Nov 9 to 15, 2026) was Saturday Nov 14 at 16:00 UTC, when 48.9% of requests to Bedrock came back with a provider error.","soWhat":"Outages come in hours, not weeks: Bedrock averaged 2.30% for the week. A fallback policy turns an hour like that into a slower answer instead of a failed one.","action":{"text":"Set up a fallback policy","href":"https://docs.requesty.ai/features/fallback-policies"},"metric":{"value":48.9,"unit":"pct","denominator":"of requests to Bedrock in the worst hour"},"footnote":"Worst hourly provider error rate by host, UTC · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"pct","rows":[{"label":"Bedrock","value":48.9,"highlight":true},{"label":"Melious","value":34.6,"highlight":false},{"label":"Sference","value":23.2,"highlight":false},{"label":"NVIDIA","value":23.1,"highlight":false},{"label":"TensorX","value":15.5,"highlight":false},{"label":"Z.ai","value":12.6,"highlight":false},{"label":"Fireworks","value":12.4,"highlight":false},{"label":"Anthropic","value":7.6,"highlight":false},{"label":"Novita","value":7.5,"highlight":false},{"label":"Mistral","value":5.2,"highlight":false}]},"posts":{"x":["Worst hour on Requesty last week: Saturday Nov 14, 16:00 UTC. 48.9% of requests to Bedrock hit a provider error. Its average for the week: 2.30%.","Outages come in hours, not weeks. Bedrock averaged 2.30% provider errors on Requesty last week, but 48.9% in its worst hour (Sat 14, 16:00 UTC).","Every host has a bad hour. Worst single hour per host on Requesty last week: Bedrock 48.9%, Melious 34.6%, Sference 23.2% (Nov 9 to 15)."],"linkedin":"The worst hour for any busy host on the Requesty gateway in week 46 (Nov 9 to 15, 2026) was Saturday Nov 14 at 16:00 UTC, when 48.9% of requests to Bedrock came back with a provider error.\n\nOutages come in hours, not weeks: Bedrock averaged 2.30% for the week. A fallback policy turns an hour like that into a slower answer instead of a failed one.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}},{"id":"warm-cache","parent":"prompt-cache","chart":"cache","topic":"Caching","title":"What a warm cache buys","lede":"Median time to first token by prompt size, for requests that read more than half their input from a prompt cache against those that did not.","short":"A warm cache started 8.2× sooner.","claim":"On the Requesty gateway in week 46 (Nov 9 to 15, 2026), deepseek-v4.1-flash prompts over 512k tokens reached the first token 8.2x sooner when more than half of the input came from cache: 2.65 s against 22 s at the median.","soWhat":"Below about 8k tokens the cache barely changes the wait. Above 32k it is the difference between an agent that feels instant and one that stalls, so keep the start of long prompts stable.","action":{"text":"Turn on auto caching","href":"https://docs.requesty.ai/features/auto-caching"},"metric":{"value":8.2,"unit":"x","denominator":"cold vs warm median time to first token, deepseek-v4.1-flash, prompts over 512k tokens"},"footnote":"Median TTFT at 128k tokens, warm bar inside the cold outline · Nov 9 to 15, 2026","figure":{"kind":"bars","unit":"ms","rows":[{"label":"deepseek-v4.1-flash","value":1072,"prev":5333,"highlight":true},{"label":"gpt-6-luna","value":4557,"prev":7196,"highlight":false},{"label":"claude-haiku-5-5","value":2619,"prev":4740,"highlight":false},{"label":"glm-5.3-flash","value":1141,"prev":7445,"highlight":false},{"label":"claude-opus-5-5","value":3735,"prev":7553,"highlight":false},{"label":"gpt-6.1-sol","value":4985,"prev":5316,"highlight":false},{"label":"deepseek-v4-flash","value":1069,"prev":7500,"highlight":false},{"label":"claude-sonnet-5-5","value":3005,"prev":5973,"highlight":false}]},"posts":{"x":["What a warm cache buys: deepseek-v4.1-flash on Requesty last week, prompts over 512k tokens, 22 s to the first token cold and 2.65 s warm. 8.2x sooner (Nov 9 to 15).","Prompt caching is a speed feature too. Over 512k tokens, deepseek-v4.1-flash started 8.2x sooner on Requesty last week when most of the prompt was cached (Nov 9 to 15).","Below 8k tokens a warm cache barely changes time to first token. Above 32k it decides it. deepseek-v4.1-flash over 512k tokens on Requesty last week: 22 s cold, 2.65 s warm."],"linkedin":"On the Requesty gateway in week 46 (Nov 9 to 15, 2026), deepseek-v4.1-flash prompts over 512k tokens reached the first token 8.2x sooner when more than half of the input came from cache: 2.65 s against 22 s at the median.\n\nBelow about 8k tokens the cache barely changes the wait. Above 32k it is the difference between an agent that feels instant and one that stalls, so keep the start of long prompts stable.\n\nThe Gateway Charts is our weekly read of what runs through the Requesty gateway: models, apps, speed and reliability, with the data behind every chart."}}]}