Compare models
Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.
Claude Haiku 4.5Qwen3 235B A22B Thinking 2507
| Claude Haiku 4.5 Anthropic | Qwen3 235B A22B Thinking 2507 Alibaba (Qwen) | |
|---|---|---|
| Pricing | ||
| Input $/1M | $1.00 | $0.30 |
| Output $/1M | $5.00 | $3.00 |
| Blended $/1M (3:1) | $2.00 | $0.975 |
| Cache read $/1M | $0.10 | — |
| Cache write $/1M | $1.25 | — |
| Quality indices | ||
| Intelligence Index | 29.6 | 19.6 |
| Coding Index | 43.9 | 22.1 |
| Math Index | — | — |
| Agentic Index | 16.4 | 3.8 |
| Speed & latency | ||
| Output speed (tok/s) | — | — |
| Time to first token | — | — |
| Time to first answer token | — | — |
| Design Arena (Elo) | ||
| Website | 1,149 | 1,080 |
| UI components | 1,142 | 985 |
| Data viz | 1,160 | 984 |
| SVG | 1,072 | — |
| 3D | 1,131 | 1,058 |
| Game dev | 1,155 | 1,016 |
| Code categories | 1,148 | 1,066 |
| ASCII art | 1,182 | — |
| Specs | ||
| Context window | 200K | 262K |
| Max output tokens | 64K | 33K |
| Input modalities | text, image, file | text |
| Tokenizer | Claude | Qwen3 |