DeepSeek V3.2-Exp scores 20.3 on the Respan Index (#95), ahead of Claude Opus 4 at 20.0 (#96). Claude Opus 4 leads on 3 of the 3 benchmarks both report. Claude Opus 4 has the larger context window (200K tokens).
| Benchmark | Claude Opus 4 | DeepSeek V3.2-Exp |
|---|---|---|
| WeirdML (Håvard Tveit Ihle) | 43.7% | 39.5% |
| LMArena Elo | 1426 | 1425 |
| Terminal-bench | 43.2% | 37.7% |
Higher is better. Each score links to where it was published.
* Estimated: no published scores in that category. How the Respan Index works