Claude Sonnet 4 and DeepSeek V3.1-Terminus tie on the Respan Index at 18.7. DeepSeek V3.1-Terminus leads on 1 of the 1 benchmarks both report. Claude Sonnet 4 has the larger context window (200K tokens).
Higher is better. Each score links to where it was published.
* Estimated: no published scores in that category. How the Respan Index works