| Benchmark | GPT-5.1 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 32.2 | 39.7 |
| AA Intelligence | 37.5 | 53.2 |
| AA-LCR | 76.7 | 81.3 |
| AA-Omniscience | 5.4 | 28.1 |
| arena_elo | 1439 | 1491 |
| arena_text_factuality | 1445 | 1466 |
| Artificial Analysis Coding Index | 49.4 | 71.3 |
| coding_arena_elo | 1421 | 1539 |
| critpt | 4.9 | 15.1 |
| Fortress | 25.7 | 12.4 |
| gdpval | 25 | 43.7 |
| GPQA Diamond | 88.1 | 89.8 |
| HLE | 28.5 | 62.1 |
| MCP Atlas | 50.1 | 88.1 |
| multichallenge | 63.4 | 75.3 |
| MultiNRC | 49 | 65.6 |
| OmniScience Accuracy | 37.7 | 52 |
| OmniScience Non-Hallucination | 48.1 | 50 |
| scicode | 43.3 | 58.2 |
| TauBench V3 - Banking | 15.9 | 31.8 |
| Terminal-Bench 2.1 | 52.4 | 80 |
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.