Compare models
Comparing 2models. Drop the URL into a doc — it's permalinked.
| Field | openai/o1 | deepseek/deepseek-reasoner |
|---|---|---|
| Provider | openai | deepseek |
| Model ID | o1 | deepseek-reasoner |
| Context | 200K | 64K |
| Max output | 100K | 8K |
| Input / 1M | $15.00 | $0.55 |
| Output / 1M | $60.00 | $2.19 |
| Cached input / 1M | $7.50 | $0.14 |
| Avg cost / 1M | $37.50 | $1.37 |
| Speed | 35 t/s | 30 t/s |
| Quality index | 85.0 | 76.0 |
| MMLU | 92.3 | 90.8 |
| GPQA | 78.0 | 71.5 |
| HumanEval | 89.0 | 86.4 |
| MATH | 94.8 | 97.3 |
| SWE-bench | 48.9 | — |
| Arena Elo | — | — |
| Tools | ✓ | — |
| Vision | ✓ | — |
| Thinking | ✓ | ✓ |
| Streaming | ✓ | ✓ |
| JSON mode | — | — |
| Structured output | ✓ | — |
| Prompt cache | — | — |
Same data, in your terminal: relay models compare o1-pro r1