Current status
- Core services operational
- All 153 monitored models operational
6 models are scheduled for deprecation (GLM-5.1, Qwen3.5-122B-A10B, Step-3.7-Flash and 3 more) — first one today
Last updated
What do the colors mean?
- Operational
- Recent health checks are passing.
- On a day in the 90-day strip: Under about 7 minutes lost that day. A short blip still shows its exact duration on hover, and still counts against the percentage.
- Degraded performance
- Some recent health checks are failing or slow. Most are still passing.
- On a day in the 90-day strip: Roughly 7 to 29 minutes lost across that day.
- Partial outage
- A substantial share of recent health checks are failing.
- On a day in the 90-day strip: Roughly 29 minutes to 2.4 hours lost across that day.
- Major outage
- The majority of recent health checks are failing. It may still be serving some traffic.
- On a day in the 90-day strip: More than about 2.4 hours lost that day, up to unavailable throughout.
- Under maintenance
- Planned work, announced in advance. Not counted against uptime.
- No data
- Not measured.
- On a day in the 90-day strip: Not measured that day. The model was newer than the window, or not covered by a health check, so it is excluded from the percentage rather than counted as either up or down.
Each bar is one day, oldest on the left, colored by how much of that day was lost rather than by the worst moment it touched, so one bad minute does not paint a whole day. Hover or focus a bar for that day's date and how long it was down.
The percentage is measured, not derived from the colors, so a day can read Operational and still report a few minutes lost. Where a model has less history than the window, the figure names the shorter basis it was actually computed over rather than implying the whole window.
“Checks failing now” is faster than the badge beside it. The badge is measured over the last hour so it does not flicker; that marker reports the current probe result, so it can appear several minutes before a badge moves.
Core services
- API api.deepinfra.comOperational90-day uptime100% over 88.5 of 90 days measured90 days agoToday
- Website deepinfra.comOperational90-day uptime99.99% over 88.5 of 90 days measured90 days agoToday
Highlighted models
- DeepSeek-V4-FlashOperational90-day uptime100% over 88.5 of 90 days measured90 days agoToday
- DeepSeek-V4-Flash-0731Operational90-day uptime100% over 60.7 of 90 days measured90 days agoToday
- DeepSeek-V4.1-FlashOperational90-day uptime100% over 20.5 of 90 days measured90 days agoToday
- gemma-4-31B-it-turboOperational90-day uptime99.11% over 88.5 of 90 days measured90 days agoToday
- GLM-5.2Operational90-day uptime99.98% over 88.5 of 90 days measured90 days agoToday
- GLM-5.3Operational90-day uptime99.99% over 32.8 of 90 days measured90 days agoToday
- GLM-5.3-FlashOperational90-day uptime100% over 34.6 of 90 days measured90 days agoToday
- gpt-oss-120bOperational90-day uptime99.82% over 88.5 of 90 days measured90 days agoToday
- MiMo-V2.6-FlashOperational90-day uptime100% over 9.1 of 90 days measured90 days agoToday
- Mistral-Nemo-Instruct-2407Operational90-day uptime100% over 88.5 of 90 days measured90 days agoToday
All monitored models
Search for a specific model to see its current status and 90-day history.
153 models monitored
- all-MiniLM-L12-v2Operational
- all-MiniLM-L6-v2Operational
- all-mpnet-base-v2Operational
- Audio8-TTS-Preview-0.6bOperational
- bge-base-en-v1.5Operational
- bge-en-iclOperational
- bge-large-en-v1.5Operational
- bge-m3Operational
- bge-m3-multiOperational
- bge-m3-multi-8kOperational
- chatterbox-multilingualOperational
- chatterbox-turboOperational
Scheduled deprecations
6 models are scheduled for deprecationfirst one today
These models are serving normally until the moment shown. After that, requests are redirected to the replacement where one is named. Times are UTC.
- GLM-5.1→ zai-org/GLM-5.3
- Qwen3.5-122B-A10B→ Qwen/Qwen3.5-397B-A17B
- Step-3.7-Flash→ XiaomiMiMo/MiMo-V2.6-Flash
- Llama-4-Maverick-17B-128E-Instruct-FP8→ google/gemma-4-31B-it-turbo
- gemma-4-E4B-it→ google/gemma-4-31B-it-turbo
- Nemotron-3-Diarization-preview→ nvidia/Nemotron-3-Diarization