What we measured
Every measured job, slowest last
Share of jobs finished, against elapsed time. 73 completions, 30-day window.
The curve is drawn from the pooled measurements, while the headline percentiles above are the median of each contributor's own figures - so the two need not line up exactly. That is deliberate: no single source can move a headline number by sending more data.
What you can plan for
0s
Upper end of the measured interval
Plan for 0s. That is the upper end of a 90% confidence interval around the p90, computed from the measurements themselves. It is deliberately not the median: being wrong toward synchronous costs twice the money, being wrong toward batch can miss a deadline, and those two mistakes are not equally expensive.
The slowest job we actually observed took 0s. That is one job, not a bound.
How much this answer is worth
Confidence: Very low
Score 0.25 · basis: exact_model
Based on a conservative upper bound.
The tail is not identified: too much of the evidence is still running or was abandoned, so any percentile would be invented. We have no observed wait to report as a lower bound.
Where the numbers come from
Each measurement is one completed batch job. The duration is end minus start, both timestamped by our server - a client can never submit a duration, and a job we stopped waiting for is recorded as abandoned, never as a fast completion.
| Newest measurement | 2026-08-29T07:58:53.000Z |
|---|---|
| Oldest in window | 2026-08-29T07:38:19.000Z |
| Figures computed | 2026-08-29T10:20:48.000Z |
| Public delay | 15 minutes |
| Percentile method | unidentified (0 voting sources) |
Timestamps are absolute and in UTC on purpose: this page may be cached, and a relative age would quietly go stale in the cache while an absolute one stays true.
Questions this page answers
- How long does gpt-4o-mini batch take at OpenAI?
- Measured over the last 30 days: half of gpt-4o-mini batch jobs at OpenAI finished within 0s, across 73 completed jobs from 3 sources. That is too few to quote a ninetieth percentile yet - it lands once 8 jobs are in - so the honest answer is the median you already have.
- How current is this gpt-4o-mini figure?
- The newest gpt-4o-mini batch job behind this page finished at 2026-08-29T07:58:53.000Z. Public figures are precomputed and delayed by 15 minutes; contributors see them at five minutes, and paid access is computed live.
Get this as JSON
Same numbers, same delay, no key required.
curl "https://batchwatch.dev/v1/curve?provider=openai&model=gpt-4o-mini"
Contributing five measurements in seven days cuts the delay to five minutes and unlocks the decision endpoints (/v1/should-i-batch, /v1/estimate-batchtime). Start on the front page.