GPT-6.2 Astra 28.43
OpenAI · Codex · 272k
- Score
- 28.43
- Rank
- 3/ 10
- Best on
- 0/ 8 tasks
- API cost
- $573
- Input tokens · 97.3% cached
- 705 M
- Output tokens
- 3.7 M
§ 01Per task
| Task | Judged | Rank | Tier | Self-best | Rounds | Hours | API cost | |
|---|---|---|---|---|---|---|---|---|
| LLaVA | 2.74× | 3 / 10 | N1 | 2.77× | 65 | 10.0 h | $114 | trace → |
| YOCO | 2.01× | 3 / 10 | N1 | 2.02× | 40 | 10.0 h | $81.4 | trace → |
| NanoChat | 1.17× | 3 / 10 | N1 | 1.17× | 17 | 10.0 h | $72.3 | trace → |
| BAGEL SFT | 1.84× | 4 / 10 | N1 | 1.93× | 18 | 10.0 h | $45.2 | trace → |
| R3 | 1.77× | 3 / 10 | N0 | 1.81× | 15 | 10.0 h | $50.0 | trace → |
| 3DGS | 3.58× | 3 / 10 | N1 | 3.54× | 57 | 10.0 h | $97.4 | trace → |
| Slime GRPO | 2.15× | 3 / 10 | N1 | 1.98× | 12 | 10.0 h | $59.1 | trace → |
| Slime OPD | 1.46× | 4 / 10 | N1 | 1.46× | 13 | 10.0 h | $54.3 | trace → |