Gemini 4 Argon on EgoSchema
86.4%Rank #1 of 10 on EgoSchema
- Coverage
- Full (500/500)
- Cost per task
- $0.0266
- Median latency
- 10.3s
- Tokens per task
- 12,930 in · 74 out
Run setting: 1 fps.
Tokens and latency
Tokens per call
| Token type | Mean | Median | 90th pct. |
|---|---|---|---|
| Input | 12,930 | 12,828 | 13,340 |
| of which cached | 61 | 0 | 0 |
| Output | 74 | 5 | 164 |
| of which reasoning | 69 | 0 | 159 |
Over 999 calls that returned an answer, including questions asked again after a failed attempt.
Latency
- Median
- 10.3s
- Mean
- 14.8s
- 90th percentile
- 21.0s
- 99th percentile
- 87.1s
Per call, end to end including file upload and processing.