Across 5 shared benchmarks, Claude Sonnet 4-20250514 (no thinking) scores higher on 3 and Gemini 3 Pro on 2. The widest gap is vectara_avg_summary_length, where Claude Sonnet 4-20250514 (no thinking) scores 145.8 against 101.9.
| Benchmark | Claude Sonnet 4-20250514 (no thinking) | Gemini 3 Pro |
|---|---|---|
| arena_vision | 1176 | 1305 |
| vectara_answer_rate | 98.6 | 99.4 |
| vectara_avg_summary_length | 145.8 | 101.9 |
| vectara_factual_consistency | 89.7 | 86.4 |
| vectara_hallucination_rate ↓ | 10.3 | 13.6 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.