OpenAI Decisions API status
How fast the OpenAI Decisions API starts answering from 6 cities, checked every 30 minutes.
- AI API
- OpenAI
- gpt-6-luna
Copy link
Last checked Every 30 minutes
Slower than usual in 2 of 6 cities
OpenAI Decisions API full response is slower than usual from Montreal: full response 291 ms vs a usual 167 ms, since 2026-10-25 21:30 UTC. 1 other city is affected too.
- Typical full response time
- 277 ms
- Last 24 hours · usually 257 ms
- Slower full response time
- 882 ms
- 95% of checks came in under this
- Failed requests
- 0
- Out of 822 requests in 24 hours
- Open incidents
- 1
- Since
OpenAI Decisions API by city
The latest full response time from each test location against its usual value. Each mark in the strip is one check.
| City | Status | Now | Usual | Last 24 hours |
|---|---|---|---|---|
 Amsterdam | Normal | 904 ms | 785 ms | |
 San Francisco | Normal | 203 ms | 121 ms | |
 Montreal | Slower than usual | 291 ms | 167 ms | |
 Singapore | Normal | 335 ms | 289 ms | |
 Tokyo | Normal | 241 ms | 225 ms | |
 Mumbai | Slower than usual | 531 ms | 344 ms |
- Normal
- Slower than usual
- Some requests failed or degraded
- No check, or ours couldn't be verified
Colours compare each city with its own usual value, not with other services or cities. In the strip, grey marks are normal checks.
Last 24 hours
Full response time for each check, as the median across all cities.
Recent incidents
All incidents →- Slower than usualMumbai, Montreal
Ongoing since · peak 531 ms, usually 344 ms
What happened →
Brief slowdowns in our checks
Two or more slow checks in a row over the last seven days that didn't last long enough to count as an incident.
- Tokyo · slower for about an hour · peak 474 ms, usually 224 ms
- Singapore · slower for about an hour · peak 608 ms, usually 289 ms
- Montreal · slower for about an hour · peak 342 ms, usually 167 ms
Show 32 more
- Montreal · slower for 3 hours · peak 446 ms, usually 166 ms
- Montreal · slower for about an hour · peak 506 ms, usually 166 ms
- Montreal · slower for 2 hours · peak 701 ms, usually 166 ms
- Montreal · slower for 90 minutes · peak 713 ms, usually 165 ms
- Montreal · slower for 2.5 hours · peak 639 ms, usually 165 ms
- Montreal · slower for 90 minutes · peak 483 ms, usually 165 ms
- Tokyo · slower for 2 hours · peak 534 ms, usually 225 ms
- San Francisco · slower for 90 minutes · peak 267 ms, usually 120 ms
- Tokyo · slower for about an hour · peak 813 ms, usually 225 ms
- Montreal · slower for 90 minutes · peak 443 ms, usually 167 ms
- Tokyo · slower for about an hour · peak 463 ms, usually 222 ms
- Montreal · slower for about an hour · peak 331 ms, usually 167 ms
- Singapore · slower for 90 minutes · peak 794 ms, usually 298 ms
- Tokyo · slower for about an hour · peak 1.08 s, usually 222 ms
- Montreal · slower for 90 minutes · peak 367 ms, usually 167 ms
- San Francisco · slower for about an hour · peak 279 ms, usually 121 ms
- Singapore · slower for about an hour · peak 512 ms, usually 298 ms
- Tokyo · slower for 2 hours · peak 499 ms, usually 216 ms
- Tokyo · slower for 90 minutes · peak 458 ms, usually 216 ms
- Montreal · slower for 90 minutes · peak 884 ms, usually 161 ms
- Montreal · slower for about an hour · peak 358 ms, usually 161 ms
- Tokyo · slower for about an hour · peak 817 ms, usually 221 ms
- Tokyo · slower for about an hour · peak 692 ms, usually 224 ms
- Singapore · slower for about an hour · peak 527 ms, usually 298 ms
- Tokyo · slower for 90 minutes · peak 655 ms, usually 222 ms
- Tokyo · slower for 90 minutes · peak 773 ms, usually 219 ms
- Tokyo · slower for 90 minutes · peak 408 ms, usually 220 ms
- Montreal · slower for about an hour · peak 306 ms, usually 161 ms
- Tokyo · slower for about an hour · peak 567 ms, usually 216 ms
- Tokyo · slower for about an hour · peak 386 ms, usually 216 ms
- San Francisco · slower for about an hour · peak 565 ms, usually 122 ms
- Tokyo · slower for 3 hours · peak 677 ms, usually 216 ms
How these checks work
- Every 30 minutes, each of our 6 test locations sends 3 requests to every service on this page.
- We compare each result with the same city's usual value over the last 7 days. A city is slower than usual above 1.5× that value (and at least 100 ms more), and degraded above 2.5×.
- If 10% or more of a city's recent requests fail, it shows "Some requests failed"; from 25% it is degraded.
- One city has to stay that way for 90 minutes before we show it. When 3 or more cities see it at the same time, we show it straight away. It takes 90 minutes of normal checks to clear.
- Each city also fetches two small files from two unrelated networks. If both are slow too, the problem is most likely on our side, and we say so instead of blaming the service.
How fast is your API around the world?
Run a free speed test from multiple cities and find out where your users are waiting. No setup, no account required.
Test my API
No account required · Takes about 30 seconds.