Compare models
Comparing 2models. Drop the URL into a doc — it's permalinked.
| Field | openai/gpt-4o | deepseek/deepseek-chat |
|---|---|---|
| Provider | openai | deepseek |
| Model ID | gpt-4o | deepseek-chat |
| Context | 128K | 164K |
| Max output | 16K | 8K |
| Input / 1M | $2.50 | $0.32 |
| Output / 1M | $10.00 | $0.89 |
| Cached input / 1M | $1.25 | $0.07 |
| Avg cost / 1M | $6.25 | $0.60 |
| Speed | 110 t/s | 50 t/s |
| Quality index | 71.0 | 72.0 |
| MMLU | 88.7 | 87.1 |
| GPQA | 53.6 | 59.1 |
| HumanEval | 90.2 | 89.0 |
| MATH | 76.6 | 90.2 |
| SWE-bench | — | — |
| Arena Elo | 1338 | — |
| Tools | ✓ | ✓ |
| Vision | ✓ | — |
| Thinking | — | — |
| Streaming | ✓ | ✓ |
| JSON mode | ✓ | ✓ |
| Structured output | ✓ | ✓ |
| Prompt cache | — | — |
Same data, in your terminal: relay models compare 4o deepseek-v3