| Benchmark | Claude Sonnet 4.5 | o4-mini |
|---|---|---|
| AA Agentic Index | 50.6 | 36.1 |
| AA Intelligence | 37.4 | 26.1 |
| AA-LCR | 68.3 | 60 |
| AA-Omniscience | -0.1 | -35.7 |
| AIME 2025 | 88 | 92.7 |
| AIR-Bench 2024 | 89.8 | 78.5 |
| arc_agi_1 | 63.7 | 58.7 |
| ARC-AGI-2 | 13.6 | 6.1 |
| Artificial Analysis Coding Index | 52.1 | 25.6 |
| browsecomp | 24.1 | 51.5 |
| critpt | 1.1 | 0.6 |
| Fortress | 17 | 21.5 |
| frontiermath_tier_4 | 4.2 | 6.3 |
| gdpval | 40.4 | 25.4 |
| GPQA Diamond | 83.4 | 81.4 |
| HLE | 19.8 | 16.5 |
| IFBench | 57.3 | 68.7 |
| LiveCodeBench | 71 | 84.5 |
| MathVista | 79.8 | 84.3 |
| MMMU | 79.6 | 81.6 |
| MMMU-Pro | 68.7 | 69.2 |
| multichallenge | 55.3 | 44.9 |
| MultiNRC | 35.8 | 22.2 |
| OmniScience Accuracy | 32.9 | 24.8 |
| OmniScience Non-Hallucination | 50.9 | 19.5 |
| scicode | 45 | 46.5 |
| SEAL VISTA | 48.8 | 51.8 |
| simplebench | 54.3 | 38.7 |
| simpleqa_verified | 23.6 | 23.9 |
| swe_bench_bash | 71.4 | 45 |
| SWE-bench Verified | 82 | 68.1 |
| TAU-bench (airline) | 70 | 49.2 |
| TAU-bench (retail) | 86.2 | 71.8 |
| Terminal-Bench Hard | 35.6 | 15.2 |
| vectara_answer_rate | 95.6 | 99.2 |
| vectara_avg_summary_length | 127.8 | 130.9 |
| vectara_factual_consistency | 88 | 81.4 |
| vectara_hallucination_rate ↓ | 12 | 18.6 |
| VisualToolBench | 6.2 | 11.1 |
| τ²-Bench Telecom (AA run) | 78.1 | 55.6 |
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.