Claude Sonnet 5.5 36.66
Anthropic · Claude Code · 1M
- Score
- 36.66
- Rank
- 2/ 10
- Best on
- 4/ 8 tasks
- API cost
- $406
- Input tokens · 99.4% cached
- 2494 M
- Output tokens
- 9.8 M
§ 01Per task
| Task | Judged | Rank | Tier | Self-best | Rounds | Hours | API cost | |
|---|---|---|---|---|---|---|---|---|
| LLaVA | 2.80× | 2 / 10 | N1 | 2.74× | 18 | 10.0 h | $39.7 | trace → |
| YOCO | 2.26× | 2 / 10 | N1 | — | 18 | 10.0 h | $43.8 | trace → |
| NanoChat | 1.21× | 1 / 10 | N1 | 1.19× | 24 | 10.0 h | $48.4 | trace → |
| BAGEL SFT | 2.21× | 1 / 10 | N1 | 2.24× | 14 | 10.0 h | $46.7 | trace → |
| R3 | 1.96× | 2 / 10 | N0 | 1.92× | 9 | 10.0 h | $37.9 | trace → |
| 3DGS | 4.44× | 2 / 10 | N1 | 4.06× | 20 | 10.0 h | $56.5 | trace → |
| Slime GRPO | 2.40× | 1 / 10 | N1 | 2.45× | 10 | 10.0 h | $58.3 | trace → |
| Slime OPD | 1.76× | 1 / 10 | N1 | 1.68× | 6 | 10.0 h | $75.2 | trace → |