AI models
Claude Opus 5.5 vs Claude Sonnet 5.5
Same criteria, side by side. Where the gap is small, it may be measurement noise.
Also searched as
Claude Opus vs Claude Sonnet
Direct answer: GPQA Diamond: Claude Opus 5.5 90.6%, Claude Sonnet 5.5 95.6% (Epoch AI). OTIS Mock AIME: Claude Opus 5.5 100%, Claude Sonnet 5.5 100% (Epoch AI). FrontierMath (tiers 1 to 3): Claude Opus 5.5 91.2%, Claude Sonnet 5.5 88.8% (Epoch AI). SimpleQA Verified: Claude Opus 5.5 72.2%, Claude Sonnet 5.5 46.5% (Epoch AI). GPQA Diamond (OpenRouter): Claude Opus 5.5 90.2%, Claude Sonnet 5.5 93% (OpenRouter Benchmarks). τ-bench Verified (airline): Claude Opus 5.5 75.3%, Claude Sonnet 5.5 78.8% (OpenRouter Benchmarks).
| Metric | Claude Opus 5.5 | Claude Sonnet 5.5 | Source |
|---|---|---|---|
| Weights ?what it is | closed | closed | no data |
| Input US$/1M ? | US$ 4.00 | US$ 2.00 | no data |
| Output US$/1M ? | US$ 20.00 | US$ 10.00 | no data |
| Context ? | 1,000,000 | 1,000,000 | no data |
| GPQA Diamond ? | 90.6% | 95.6% | Epoch AI |
| OTIS Mock AIME ? | 100% | 100% | Epoch AI |
| FrontierMath (tiers 1 to 3) ? | 91.2% | 88.8% | Epoch AI |
| SimpleQA Verified ? | 72.2% | 46.5% | Epoch AI |
| GPQA Diamond (OpenRouter) ? | 90.2% | 93% | OpenRouter Benchmarks |
| τ-bench Verified (airline) ? | 75.3% | 78.8% | OpenRouter Benchmarks |
Compare with
Sources
- Epoch AI, CC BY 4.0, latest version we checked.
- OpenRouter Benchmarks, CC BY 4.0, latest version we checked.
- OpenRouter, dados públicos da API; exports de benchmarks e rankings em CC BY 4.0, latest version we checked.
Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.