GPT-6 Astra on UCF101-AD
61.4%Rank #2 of 9 on UCF101-AD
- Coverage
- Full (4224/4224)
- Cost per task
- $0.0221
- Median latency
- 2.9s
- Tokens per task
- 2,109 in · 20 out
Run setting: 1 fps.
Tokens and latency
Tokens per call
| Token type | Mean | Median | 90th pct. |
|---|---|---|---|
| Input | 2,109 | 1,763 | 2,755 |
| of which cached | 1 | 0 | 0 |
| Output | 20 | 4 | 49 |
| of which reasoning | 14 | 0 | 39 |
Over 4,224 calls that returned an answer, including questions asked again after a failed attempt.
Latency
- Median
- 2.9s
- Mean
- 3.2s
- 90th percentile
- 4.8s
- 99th percentile
- 7.6s
Per call, end to end including file upload and processing.