Previous edition: Operational metrics per provider, April 2026
Operational metrics per provider, October 2026
Operational metrics per provider, October 2026
Switch metrics. Hover any row to see all three at once.
Share of requests that completed without an upstream error.Source: Requesty production gateway, October 2026.
How do the largest providers compare on success rate, streaming and caching? In October 2026 Parasail had the highest success rate at 100% and DeepInfra the lowest at 91% among the 15 busiest providers. Melious streamed the largest share of requests at 70%.
Why it mattersSuccess rate, streaming share and cache hit rate together describe how a provider behaves under production load. A low success rate means more retries or fallbacks; caching and streaming change cost and perceived latency.
Period
Oct 2026
Updated
November 11, 2026
ID
ops-metrics-october-2026
§ 01
Key findings
- 01Highest success rate: Parasail at 100%.
- 02Lowest success rate: DeepInfra at 91%.
- 03Most streaming: Melious at 70%.
§ 01b
Questions this answers
- Which AI provider has the highest success rate?
- Which providers stream the most requests?
- How do providers compare on caching?
§ 02
Data
| Provider | Success rate(percent) | Streaming(percent) | Cache hit(percent) |
|---|---|---|---|
| Vertex (Gemini) | 96.52% | 10.14% | 50.52% |
| Azure | 91.61% | 31.82% | 71.14% |
| Typesafe | 99.44% | 0.04% | 0.00% |
| TensorX | 91.60% | 9.15% | 66.09% |
| OpenAI | 98.41% | 10.87% | 76.58% |
| DeepInfra | 91.31% | 9.62% | 75.98% |
| Nebius | 99.43% | 0.52% | 50.87% |
| Bedrock | 92.46% | 27.82% | 82.41% |
| Sference | 95.72% | 37.48% | 82.21% |
| Anthropic | 93.01% | 43.48% | 80.36% |
| DeepSeek | 96.27% | 31.29% | 88.53% |
| Vertex (Claude) | 92.74% | 47.60% | 82.28% |
| Novita | 92.52% | 11.60% | 84.53% |
| Melious | 94.38% | 69.62% | 90.05% |
| Parasail | 99.50% | 0.18% | 62.36% |
§ 03
Cite as
APA
Click to copy
BibTeX
Click to copy
ID: ops-metrics-october-2026·Updated November 11, 2026·Period Oct 2026