Provider reliability · tracked daily
Which provider is most reliable for each model?
Uptime and latency, per provider, per model - measured directly from OpenRouter's own endpoint data. The uptime figure is OpenRouter's own trailing-30-minute metric, captured once daily by our sync (not a live/real-time feed) - a daily snapshot of that window, not this instant. A model's benchmark score doesn't tell you whether you can actually depend on getting a response; this does.
Which host is most reliable overall?
Azure leads among hosts serving 3+ tracked models, at 100.0% median uptime across 21 models. Inverted from the table above: instead of "which host is best for this model," this ranks each host by measured reliability across every model it serves.
| # | Provider | Models served | Median uptime | Avg uptime | Peak throughput |
|---|---|---|---|---|---|
| 01 | Azure | 21 | 100.0% | 98.0% | 128 tok/s |
| 02 | Google AI Studio | 10 | 100.0% | 100.0% | 144 tok/s |
| 03 | Anthropic | 9 | 100.0% | 100.0% | 63 tok/s |
| 04 | BaseTen | 9 | 100.0% | 99.7% | 171 tok/s |
| 05 | Io Net | 7 | 100.0% | 93.3% | 108 tok/s |
| 06 | Baidu | 6 | 100.0% | 99.8% | 80 tok/s |
| 07 | Claude Platform on AWS | 6 | 100.0% | 100.0% | 58 tok/s |
| 08 | Meta | 4 | 100.0% | 100.0% | 126 tok/s |
| 09 | Ambient | 2 | 100.0% | 100.0% | 32 tok/s |
| 10 | Decart | 2 | 100.0% | 100.0% | 182 tok/s |
| 11 | Nex AGI | 2 | 100.0% | 100.0% | 94 tok/s |
| 12 | Poolside | 2 | 100.0% | 100.0% | 114 tok/s |
| 13 | Tencent | 2 | 100.0% | 100.0% | 40 tok/s |
| 14 | Alibaba | 12 | 100.0% | 99.7% | 84 tok/s |
| 15 | Mistral | 2 | 100.0% | 100.0% | 117 tok/s |
| 16 | Amazon Bedrock | 19 | 100.0% | 100.0% | 128 tok/s |
| 17 | Stealth | 2 | 100.0% | 100.0% | 26 tok/s |
| 18 | ModelRun | 3 | 100.0% | 97.2% | 107 tok/s |
| 19 | OpenAI | 10 | 100.0% | 99.8% | 98 tok/s |
| 20 | AtlasCloud | 13 | 100.0% | 99.6% | 94 tok/s |
| 21 | DeepSeek | 4 | 100.0% | 99.6% | 78 tok/s |
| 22 | Relace | 4 | 100.0% | 98.1% | 78 tok/s |
| 23 | AkashML | 6 | 100.0% | 99.8% | 109 tok/s |
| 24 | Sail Research | 6 | 100.0% | 99.8% | 66 tok/s |
| 25 | CoreWeave | 12 | 100.0% | 99.6% | 158 tok/s |
| 26 | Groq | 3 | 99.9% | 99.9% | 453 tok/s |
| 27 | Darkbloom | 4 | 99.9% | 99.9% | 33 tok/s |
| 28 | Wafer | 7 | 99.9% | 99.7% | 90 tok/s |
| 29 | xAI | 2 | 99.9% | 99.9% | 62 tok/s |
| 30 | Friendli | 5 | 99.9% | 99.8% | 109 tok/s |
| 31 | Z.AI | 4 | 99.9% | 99.9% | 66 tok/s |
| 32 | Modal | 4 | 99.9% | 99.9% | 136 tok/s |
| 33 | Upstage | 1 | 99.9% | 99.9% | 39 tok/s |
| 34 | Near AI | 1 | 99.9% | 99.9% | 24 tok/s |
| 35 | Novita | 21 | 99.9% | 95.4% | 104 tok/s |
| 36 | Mancer 2 | 3 | 99.9% | 99.3% | 66 tok/s |
| 37 | Together | 12 | 99.9% | 99.2% | 156 tok/s |
| 38 | Moonshot AI | 1 | 99.8% | 99.8% | 24 tok/s |
| 39 | DigitalOcean | 11 | 99.8% | 99.5% | 52 tok/s |
| 40 | Parasail | 16 | 99.8% | 98.3% | 129 tok/s |
| 41 | Phala | 14 | 99.8% | 92.5% | 111 tok/s |
| 42 | Ionstream | 3 | 99.8% | 99.4% | 66 tok/s |
| 43 | SiliconFlow | 13 | 99.8% | 96.4% | 88 tok/s |
| 44 | 23 | 99.7% | 96.6% | 168 tok/s | |
| 45 | Xiaomi | 4 | 99.7% | 99.7% | 39 tok/s |
| 46 | Crusoe | 5 | 99.7% | 98.0% | 214 tok/s |
| 47 | DekaLLM | 4 | 99.7% | 99.1% | 44 tok/s |
| 48 | DeepInfra | 30 | 99.7% | 97.9% | 132 tok/s |
| 49 | OpenInference | 4 | 99.7% | 99.5% | 32 tok/s |
| 50 | StepFun | 1 | 99.7% | 99.7% | 36 tok/s |
| 51 | Inceptron | 3 | 99.7% | 99.6% | 74 tok/s |
| 52 | Reka | 5 | 99.6% | 99.2% | 55 tok/s |
| 53 | Venice | 15 | 99.6% | 98.6% | 129 tok/s |
| 54 | StreamLake | 10 | 99.5% | 98.6% | 77 tok/s |
| 55 | Fireworks | 10 | 99.5% | 69.5% | 129 tok/s |
| 56 | Makora | 6 | 99.4% | 98.1% | 112 tok/s |
| 57 | Krea | 1 | 99.3% | 99.3% | 114 tok/s |
| 58 | Minimax | 2 | 99.2% | 99.2% | 68 tok/s |
| 59 | GMICloud | 13 | 99.2% | 98.4% | 82 tok/s |
| 60 | Morph | 8 | 98.1% | 96.0% | 35 tok/s |
| 61 | NextBit | 4 | 98.1% | 97.9% | 56 tok/s |
| 62 | SambaNova | 5 | 97.2% | 96.7% | 387 tok/s |
| 63 | InferenceNet | 3 | 97.1% | 97.5% | 68 tok/s |
| 64 | Chutes | 3 | 96.6% | 96.6% | 30 tok/s |
| 65 | Cloudflare | 7 | 95.8% | 96.1% | 54 tok/s |
| 66 | Mara | 4 | 93.6% | 93.6% | 179 tok/s |
| 67 | Nebius | 2 | 61.2% | 61.2% | 79 tok/s |
| 68 | Cerebras | 2 | 50.0% | 50.0% | 622 tok/s |