Claude Sonnet 4 scores 18.7 on the Respan Index (#99), ahead of Qwen3 235B-A22B Instruct 2507 at 18.5 (#101). They split the 2 benchmarks both report evenly. Qwen3 235B-A22B Instruct 2507 is 15x cheaper per token at list price. Qwen3 235B-A22B Instruct 2507 has the larger context window (262K tokens).
| Benchmark | Claude Sonnet 4 | Qwen3 235B-A22B Instruct 2507 |
|---|---|---|
| WeirdML (Håvard Tveit Ihle) | 46.1% | 38.7% |
| LMArena Elo | 1402 | 1422 |
Higher is better. Each score links to where it was published.
* Estimated: no published scores in that category. How the Respan Index works