| Benchmark | Claude Fable 5 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 56.6 | 39.7 |
| AA Intelligence | 62.1 | 53.2 |
| AA-LCR | 76.7 | 81.3 |
| AA-Omniscience | 43.3 | 28.1 |
| arena_elo | 1508 | 1491 |
| arena_text_factuality | 1486 | 1466 |
| Artificial Analysis Coding Index | 76.5 | 71.3 |
| coding_arena_elo | 1626 | 1539 |
| critpt | 28.6 | 15.1 |
| deepsearchqa_f1 | 94.2 | 84.9 |
| DeepSWE 1.1 | 70 | 53.3 |
| Finance Agent v2 | 56.3 | 57.2 |
| gdpval | 61.9 | 43.7 |
| GPQA Diamond | 92.6 | 89.8 |
| healthbench_professional | 66 | 59.3 |
| HLE | 64.5 | 62.1 |
| JobBench | 57.4 | 54.7 |
| livebench_agentic_coding | 62.2 | 58.5 |
| livebench_coding | 86 | 77.2 |
| livebench_data_analysis | 80.5 | 72.5 |
| livebench_instruction_following | 75.8 | 69.6 |
| livebench_language | 90.7 | 74.3 |
| livebench_math | 96 | 87.1 |
| livebench_reasoning | 89.7 | 87.7 |
| MCP Atlas | 84.7 | 88.1 |
| OmniScience Accuracy | 65.3 | 52 |
| OmniScience Non-Hallucination | 36.4 | 50 |
| OSWorld-Verified | 85 | 80.8 |
| scicode | 60.2 | 58.2 |
| SWE-bench Pro | 80 | 61.5 |
| TauBench V3 - Banking | 38.1 | 31.8 |
| Terminal-Bench 2.1 | 88 | 80 |
| toolathlon | 61.7 | 75.6 |
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.