| Benchmark | Claude Sonnet 5 | Muse Spark |
|---|---|---|
| AA Agentic Index | 49.7 | 28.7 |
| AA Intelligence | 55.3 | 44.3 |
| AA-LCR | 77 | 77 |
| AA-Omniscience | 16.4 | 7.2 |
| arena_elo | 1463 | 1488 |
| arena_text_factuality | 1453 | 1467 |
| Artificial Analysis Coding Index | 71.5 | 58.6 |
| coding_arena_elo | 1540 | 1540 |
| critpt | 16.9 | 11.3 |
| gdpval | 54.8 | 32.3 |
| GPQA Diamond | 91.1 | 89.5 |
| HLE | 57.4 | 58.4 |
| MMMU-Pro | 77.3 | 80.5 |
| OmniScience Accuracy | 40 | 49.6 |
| OmniScience Non-Hallucination | 60.6 | 26.8 |
| scicode | 53.6 | 51.5 |
| simpleqa_verified | 25 | 66.3 |
| SWE-bench Pro | 63.2 | 55 |
| SWE-bench Verified | 85.2 | 77.4 |
| TauBench V3 - Banking | 37.3 | 19.6 |
| Terminal-Bench 2.0 | 80.4 | 59 |
| Terminal-Bench 2.1 | 80.5 | 62.2 |
| τ³-Bench | 28.2 | 19.6 |
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.