- Up: 95% or more served
- Degraded: 85% to 95%
- Down: below 85%
- No data
Every model, hour by hour
The last 72 hours, one cell per hour, and the share of requests served over those 3 days. Hatched hours are from llm.buzz's test requests, sent where customer traffic is too thin to publish.
- Anthropic: Claude Fable 5Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Fable 5.1Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Haiku 5.5Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Opus 4.7Answeringllm.buzzAnthropic98.6%test
- Anthropic: Claude Opus 4.8Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Opus 5Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Opus 5.5Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Sonnet 5Answeringllm.buzzAnthropic100%test
- Anthropic: Claude Sonnet 5.5Answeringllm.buzzAnthropic100%test
- DeepSeek: DeepSeek V4 Flash 0731AnsweringScalewayDeepInfraPrime Intellect100%test
- DeepSeek: DeepSeek V4 Flash Vision ExpAnsweringllm.buzzDeepInfra100%test
- DeepSeek: DeepSeek V4 Pro 0813Answeringllm.buzzDeepInfra100%test
- DeepSeek: DeepSeek V4.1 FlashAnsweringPrime IntellectDeepInfraScaleway100%test
- DeepSeek: DeepSeek V4.1 Flash 1017Answeringllm.buzzDeepInfraPrime Intellect100%test
- Google: Gemma 4 26B A4BAnsweringDeepInfraScalewayNovita100%test
- Google: Gemma 4 31BAnsweringllm.buzzDeepInfra100%test
- Mistral: Mistral Medium 3.5AnsweringScaleway100%test
- Mistral: Mistral Small 3.2 24BAnsweringllm.buzzScaleway100%test
- Moondream: Moondream 3 PreviewAnsweringPrime Intellect98.6%test
- Moondream: Moondream 3.1AnsweringPrime Intellect100%test
- MoonshotAI: Kimi K3Answeringllm.buzzDeepInfraMoonshot AIPrime IntellectNovitaAlibaba100%test
- MoonshotAI: Kimi K3.1 ProAnsweringMoonshot AI97.2%test
- NVIDIA: Nemotron 3 SuperAnsweringDeepInfra52.1%test
- OpenAI: GPT-4.1Answeringllm.buzzOpenAI100%test
- OpenAI: GPT-4.1 MiniAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-4o-miniAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5 MiniAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5 NanoAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5.4Answeringllm.buzzOpenAI100%test
- OpenAI: GPT-5.4 MiniAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5.4 NanoAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5.5Answeringllm.buzzOpenAI100%test
- OpenAI: GPT-5.6 LunaAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5.6 SolAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-5.6 TerraAnsweringllm.buzzOpenAI98.6%test
- OpenAI: GPT-6 AstraAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-6 LunaAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-6 SolAnsweringllm.buzzOpenAI100%test
- OpenAI: GPT-6.1 SolAnsweringllm.buzzOpenAI100%test
- OpenAI: gpt-oss-120bAnsweringllm.buzzNovitaScaleway100%test
- Qwen: Qwen3.6 35B A3BAnsweringPrime IntellectDeepInfraScaleway100%test
- Qwen: Qwen3.8 27BAnsweringDeepInfraScalewayPrime IntellectNovita100%test
- Qwen: Qwen3.8 FlashAnsweringllm.buzzNovitaAlibabaPrime Intellect100%test
- Qwen: Qwen3.8 Max (0902)AnsweringAlibaba100%test
- Qwen: Qwen3.8 Omni FlashAnsweringAlibaba98.6%test
- Surogate: Surogate 3.7 FlashAnsweringSimpleminds EUSurogate Spark100%test
- Xiaomi: MiMo-V2.6-FlashAnsweringDeepInfra97.3%test
- Xiaomi: MiMo-V2.6-ProAnsweringllm.buzzDeepInfra97.1%test
- Z.ai: GLM 5.3Answeringllm.buzzPrime IntellectAlibabaScalewayNovita100%test
- Z.ai: GLM 5.3 FlashAnsweringllm.buzzPrime IntellectDeepInfraNovita100%test
- hexgrad: Kokoro 82MNot measuredDeepInfranot measured
- OpenAI: Whisper Large V3Not measuredDeepInfraScalewaynot measured
- OpenAI: Whisper Large V3 TurboNot measuredDeepInfranot measured
- Up: 95% or more served
- Degraded: 85% to 95%
- Down: below 85%
- No data
- From llm.buzz's test requests
Failover is built inWhen a provider fails, llm.buzz sends the request to the model's next provider.