DeepSeek: R1 wins on more metrics (6 of 8), but the right pick depends on what you optimize for — see the breakdown below.
| Metric | Anthropic: Claude 3.5 Haiku | DeepSeek: R1 |
|---|---|---|
| Intelligence Index | 18.7 | 18.8✓ |
| Coding Index | 10.7 | 15.9✓ |
| GPQA Diamond | 41% | 71%✓ |
| Design Arena Elo | — | — |
| Speed (tokens/sec) | 40 | 72✓ |
| Latency | 755ms✓ | 1.4s |
| Input price /M | $0.800 | $0.700✓ |
| Output price /M | $4.00 | $2.50✓ |
| Context window | 200K✓ | 164K |
| Capabilities | ToolsVision | ReasoningToolsJSON |