The Gateway Charts · Week 45, 2026
Last week in AI. Who climbed. Who slipped.
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.
- 01 / Rankings25.9%deepseek-v4.1-flash held #1 again.
- 02 / Apps33.1%Coding agents sent a third of tokens.
- 03 / Usage68.3%Most teams used more than one model.
- 04 / Usage87.5%Most requests never set a temperature.
- 05 / Speed2.3×claude-sonnet-5 started 2.3× sooner.
- 06 / Reliability86.5%Most failed requests still got served.
- 07 / Caching77.1%77.1% of input came from cache.
01 Rankings
deepseek-v4.1-flash held #1 again.
deepseek-v4.1-flash held #1 on the Requesty gateway in week 45 (Nov 2 to 8, 2026) with 25.9% of tokens, down 3.93 points on the week before.
- Rank 1. =deepseek-v4.1-flash=DeepSeek25.9%−3.93 pts
- Rank 2. =claude-opus-5-5=Anthropic10.1%+2.27 pts
- Rank 3. =gpt-6-luna=OpenAI8.39%+0.84 pts
- Rank 4. 1Up 1 from #5glm-5.3-flash1Up 1 from #5Z.ai6.02%+1.71 pts
- Rank 5. 1Down 1 from #4gpt-6.1-sol1Down 1 from #4OpenAI5.43%−0.58 pts
- Rank 6. =claude-sonnet-5-5=Anthropic5.28%+1.39 pts
- Rank 7. 1Up 1 from #8deepseek-v4-flash1Up 1 from #8DeepSeek4.29%+0.84 pts
- Rank 8. 1Down 1 from #7claude-haiku-5-51Down 1 from #7Anthropic3.56%−0.06 pts
- Rank 9. 1Up 1 from #10nemotron-3-ultra-550b-a55b1Up 1 from #10NVIDIA2.39%−0.46 pts
- Rank 10. 1Up 1 from #11gpt-5.6-luna1Up 1 from #11OpenAI2.36%−0.02 pts
- This week
- #2 · 10.1%
- Last week
- #2 · 7.81%
- Change
- +2.27 pts
- Peak
- #1
- Weeks on chart
- 7 of 12
Data
| Rank | Last week | Name | Share | Change |
|---|---|---|---|---|
| 1 | 1 | deepseek-v4.1-flash | 25.9% | −3.93 pts |
| 2 | 2 | claude-opus-5-5 | 10.1% | +2.27 pts |
| 3 | 3 | gpt-6-luna | 8.39% | +0.84 pts |
| 4 | 5 | glm-5.3-flash | 6.02% | +1.71 pts |
| 5 | 4 | gpt-6.1-sol | 5.43% | −0.58 pts |
| 6 | 6 | claude-sonnet-5-5 | 5.28% | +1.39 pts |
| 7 | 8 | deepseek-v4-flash | 4.29% | +0.84 pts |
| 8 | 7 | claude-haiku-5-5 | 3.56% | −0.06 pts |
| 9 | 10 | nemotron-3-ultra-550b-a55b | 2.39% | −0.46 pts |
| 10 | 11 | gpt-5.6-luna | 2.36% | −0.02 pts |
02 Apps
Coding agents sent a third of tokens.
Five coding agents (OpenCode, Claude Code, Pi, DeepSeek Harness and GitHub Copilot) sent 33.1% of tokens through the Requesty gateway in week 45 (Nov 2 to 8, 2026).
- OpenAI Python SDKSDK or framework20.6%+1.09 pts
- OpenCodeCoding agent13.5%−1.59 pts
- Claude CodeCoding agent11.5%+0.71 pts
- Vercel AI SDKSDK or framework9.36%+0.83 pts
- PiCoding agent3.54%−0.56 pts
- DeepSeek HarnessCoding agent2.72%+0.66 pts
- Anthropic SDKSDK or framework2.70%+1.50 pts
- OpenAI SDK (other)SDK or framework2.44%−0.55 pts
- LiteLLMSDK or framework1.93%+0.50 pts
- GitHub CopilotCoding agent1.81%−0.90 pts
- This week
- 13.5%
- Last week
- 15.1%
- Change
- −1.59 pts
- Place in this chart
- #2 of 10
Data
| Name | This week | Last week |
|---|---|---|
| OpenAI Python SDK | 20.6% | 19.5% |
| OpenCode | 13.5% | 15.1% |
| Claude Code | 11.5% | 10.8% |
| Vercel AI SDK | 9.4% | 8.5% |
| Pi | 3.5% | 4.1% |
| DeepSeek Harness | 2.7% | 2.1% |
| Anthropic SDK | 2.7% | 1.2% |
| OpenAI SDK (other) | 2.4% | 3.0% |
| LiteLLM | 1.9% | 1.4% |
| GitHub Copilot | 1.8% | 2.7% |
03 Usage
Most teams used more than one model.
68.3% of active teams on the Requesty gateway used two or more models in week 45 (Nov 2 to 8, 2026); 59.7% used two or more labs and 61.7% two or more providers.
- Models
- 68.3%
- Labs
- 59.7%
- Providers
- 61.7%
Data
| 1 | 2 | 3 | 4 to 5 | 6 or more | Two or more | |
|---|---|---|---|---|---|---|
| Models | 31.7% | 17.7% | 13.3% | 18.2% | 19.0% | 68.3% |
| Labs | 40.3% | 23.6% | 16.7% | 14.1% | 5.2% | 59.7% |
| Providers | 38.3% | 23.8% | 14.8% | 12.7% | 10.5% | 61.7% |
04 Usage
Most requests never set a temperature.
87.5% of chat requests through the Requesty gateway in week 45 (Nov 2 to 8, 2026) never set a temperature; 77.6% left reasoning effort unset and 94.4% sent no response format.
Data
| Parameter | Value sent | Share of chat requests |
|---|---|---|
| temperature | Not set | 87.5% |
| temperature | Set to 0 | 8.4% |
| temperature | Set to 1 | 2.2% |
| temperature | Something else | 1.9% |
| reasoning_effort | Not set | 77.6% |
| reasoning_effort | On, no level | 16.7% |
| reasoning_effort | Low to max | 4.8% |
| reasoning_effort | Off or other | 0.9% |
| response_format | Free text | 94.4% |
| response_format | json_object | 2.6% |
| response_format | json_schema | 1.4% |
| response_format | Other | 1.6% |
05 Speed
claude-sonnet-5 started 2.3× sooner.
claude-sonnet-5 reached its first token 2.3x sooner on the Requesty gateway in week 45 (Nov 2 to 8, 2026): a median of 1.69 s, down from 3.8 s the week before, on streaming requests.
- claude-sonnet-53.8 s1.69 s2.3× sooner
- claude-opus-54.3 s3.1 s1.4× sooner
- gpt-6-luna1.87 s1.52 s1.2× sooner
- claude-haiku-5-52.64 s2.19 s1.2× sooner
- claude-sonnet-4-61.18 s992 ms1.2× sooner
- Last week
- 3.8 s
- This week
- 1.69 s
- Change
- 2.27× sooner
- Relative
- −56% Faster
Data
| Model | Last week | This week | Change |
|---|---|---|---|
| claude-sonnet-5 | 3.8 s | 1.69 s | 2.3× sooner |
| claude-opus-5 | 4.3 s | 3.1 s | 1.4× sooner |
| gpt-6-luna | 1.87 s | 1.52 s | 1.2× sooner |
| claude-haiku-5-5 | 2.64 s | 2.19 s | 1.2× sooner |
| claude-sonnet-4-6 | 1.18 s | 992 ms | 1.2× sooner |
06 Reliability
Most failed requests still got served.
In week 45 (Nov 2 to 8, 2026), 86.5% of requests on a fallback policy whose first attempt hit a provider error were still served by another host on the Requesty gateway; 94.5% of those rescues served the same model.
- Fallback policy
- 86.5%
- First host failed, still served
- Latency policy
- 82.5%
- First host failed, still served
- Same model served
- 94.5%
- Of rescues, same model on another host
- Served by another host
- 86.5%
- Still failed
- 13.5%
- One dot
- 0.5% of these requests
Data
| Measure | Value |
|---|---|
| Served by another host | 86.5% |
| Still failed | 13.5% |
| Fallback policy | 86.5% |
| Latency policy | 82.5% |
| Same model served | 94.5% |
07 Caching
77.1% of input came from cache.
77.1% of input tokens through the Requesty gateway in week 45 (Nov 2 to 8, 2026) were read from a prompt cache; models read 46 input tokens for every output token.
- Input read from cache
- 77.1%
- Against the week before
- −1.4 pts
- 12-week average
- 76.8%
Data
| Week of | Input read from cache |
|---|---|
| 2026-08-17 | 80.4% |
| 2026-08-24 | 82.2% |
| 2026-08-31 | 77.3% |
| 2026-09-07 | 72.5% |
| 2026-09-14 | 75.5% |
| 2026-09-21 | 74.0% |
| 2026-09-28 | 76.2% |
| 2026-10-05 | 75.7% |
| 2026-10-12 | 74.8% |
| 2026-10-19 | 77.5% |
| 2026-10-26 | 78.5% |
| 2026-11-02 | 77.1% |
Vital signs
How the traffic behaved
Week 45Full rankings
Models · 20Apps · 12Labs · 10ShowHide
| Rank | Move | Model | Share | 12 weeks | This week | Change | Peak |
|---|---|---|---|---|---|---|---|
| 1 | = | deepseek-v4.1-flashDeepSeek | 25.9% | −3.93 pts | #1 | ||
| 2 | = | claude-opus-5-5Anthropic | 10.1% | +2.27 pts | #1 | ||
| 3 | = | gpt-6-lunaOpenAI | 8.39% | +0.84 pts | #3 | ||
| 4 | 1Up 1 from #5 | glm-5.3-flashZ.ai | 6.02% | +1.71 pts | #2 | ||
| 5 | 1Down 1 from #4 | gpt-6.1-solOpenAI | 5.43% | −0.58 pts | #4 | ||
| 6 | = | claude-sonnet-5-5Anthropic | 5.28% | +1.39 pts | #6 | ||
| 7 | 1Up 1 from #8 | deepseek-v4-flashDeepSeek | 4.29% | +0.84 pts | #1 | ||
| 8 | 1Down 1 from #7 | claude-haiku-5-5Anthropic | 3.56% | −0.06 pts | #7 | ||
| 9 | 1Up 1 from #10 | nemotron-3-ultra-550b-a55bNVIDIA | 2.39% | −0.46 pts | #8 | ||
| 10 | 1Up 1 from #11 | gpt-5.6-lunaOpenAI | 2.36% | −0.02 pts | #2 | ||
| 11 | 2Down 2 from #9 | claude-sonnet-5Anthropic | 2.03% | −1.25 pts | #3 | ||
| 12 | 1Up 1 from #13 | nemotron-3.5-lightning-30b-a3bNVIDIA | 1.54% | +0.13 pts | #12 | ||
| 13 | 1Up 1 from #14 | gpt-6.1-terraOpenAI | 1.35% | −0.02 pts | #13 | ||
| 14 | 2Up 2 from #16 | gemini-3.8-flashGoogle | 1.30% | +0.36 pts | #7 | ||
| 15 | = | glm-5.3Z.ai | 1.21% | −0.02 pts | #7 | ||
| 16 | 2Up 2 from #18 | ling-3.1-flashOther | 0.93% | +0.12 pts | #16 | ||
| 17 | 8Up 8 from #25 | claude-opus-5Anthropic | 0.74% | +0.31 pts | #5 | ||
| 18 | 1Down 1 from #17 | deepseek-v4-proDeepSeek | 0.73% | −0.20 pts | #4 | ||
| 19 | 9Up 9 from #28 | gemini-3.5-flash-liteGoogle | 0.57% | +0.27 pts | #17 | ||
| 20 | 3Up 3 from #23 | claude-sonnet-4-6Anthropic | 0.53% | +0.06 pts | #10 |
| Rank | Move | App | Share | 12 weeks | This week | Change | Peak |
|---|---|---|---|---|---|---|---|
| 1 | = | OpenAI Python SDKSDK or framework | 20.6% | +1.09 pts | #1 | ||
| 2 | = | OpenCodeCoding agent | 13.5% | −1.59 pts | #1 | ||
| 3 | = | Claude CodeCoding agent | 11.5% | +0.71 pts | #2 | ||
| 4 | = | Vercel AI SDKSDK or framework | 9.36% | +0.83 pts | #3 | ||
| 5 | = | PiCoding agent | 3.54% | −0.56 pts | #4 | ||
| 6 | 2Up 2 from #8 | DeepSeek HarnessCoding agent | 2.72% | +0.66 pts | #6 | ||
| 7 | 3Up 3 from #10 | Anthropic SDKSDK or framework | 2.70% | +1.50 pts | #7 | ||
| 8 | 2Down 2 from #6 | OpenAI SDK (other)SDK or framework | 2.44% | −0.55 pts | #2 | ||
| 9 | = | LiteLLMSDK or framework | 1.93% | +0.50 pts | #9 | ||
| 10 | 3Down 3 from #7 | GitHub CopilotCoding agent | 1.81% | −0.90 pts | #5 | ||
| 11 | = | Kilo CodeCoding agent | 1.27% | +0.42 pts | #7 | ||
| 12 | 1Up 1 from #13 | LangChainSDK or framework | 0.28% | +0.04 pts | #11 |
| Rank | Move | Lab | Share | 12 weeks | This week | Change | Peak |
|---|---|---|---|---|---|---|---|
| 1 | = | DeepSeek | 32.0% | −3.10 pts | #1 | ||
| 2 | = | Anthropic | 23.8% | +2.74 pts | #1 | ||
| 3 | = | OpenAI | 20.3% | +0.30 pts | #2 | ||
| 4 | = | Z.ai | 7.76% | +1.51 pts | #2 | ||
| 5 | 1Up 1 from #6 | 4.14% | +0.39 pts | #4 | |||
| 6 | 1Down 1 from #5 | NVIDIA | 4.02% | −0.29 pts | #5 | ||
| 7 | = | MiniMax | 1.11% | −1.01 pts | #7 | ||
| 8 | 1Up 1 from #9 | Xiaomi | 0.96% | −0.09 pts | #6 | ||
| 9 | 1Up 1 from #10 | Moonshot | 0.87% | −0.17 pts | #6 | ||
| 10 | 2Down 2 from #8 | Mistral | 0.70% | −0.36 pts | #7 |
Method
How we count
Week 45 covers Nov 2 to 8, 2026 (UTC), compared with Oct 26 to Nov 1, 2026.
- 01Aggregated from the Requesty production gateway for the week (Monday to Sunday, UTC), compared with the week before and the ten weeks before that.
- 02We publish shares, rates, ranks and latency only: no volumes, no customer data, no prompt content.
- 03In 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.
- 04An entry is named only when at least 20 teams used it and no single team is more than 60% of it.
- 05Latency 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.
- 06Internal test traffic, synthetic load and abusive accounts are excluded.
- Source
- Week 45 seed: the one-off Gateway Charts report (pre-pipeline) · scripts/weekly/seed/seed-2026-w45.py
For agents
Read this week as data
- Markdown
- /data/weekly/2026-w45/data.md
- Latest
- /data/weekly/latest.json
- llms.txt
- /data/weekly/llms.txt
# Any week URL returns Markdown when asked for it
curl -H "Accept: text/markdown" \
https://www.requesty.ai/data/weekly/2026-w45Next · Week 46
