aiaimodel.cards
← All benchmarks

SWE-bench Verified (hard subset, >1 h tasks)

Fixing real issues in real open-source repositories; a patch counts only if the project's own tests pass. Variants differ in task set and languages.

Your workloadRAG over docs · 12K in · 600 out · 50K requests/mo · 30% cached
  1. 1Claude Sonnet 4.5 · Anthropicbest value
    45.3%

“Best value”: no other model here reports a higher score at a lower cost for your workload. Scores are self-reported by each maker with different settings, so treat close calls as ties.

How each maker reported it

  • Claude Sonnet 4.5 §9.2–9.3, p.130–148

    The card has no general capability table; these are the numeric Responsible Scaling Policy capability results it reports. Helpful-only variants and extended thinking were used where they scored higher. Opus SWE-bench (hard) scores are on 42 problems, Sonnet 4.5 on 45.

Only each model's own maker's number is shown; scores a maker reports for rival models are left out.