| Benchmark | Muse Spark | Qwen3.6 Plus |
|---|---|---|
| AA Agentic Index | 28.7 | 29 |
| AA Intelligence | 44.3 | 40.5 |
| AA-LCR | 77 | 72.3 |
| AA-Omniscience | 7.2 | 2.6 |
| arena_elo | 1488 | 1444 |
| arena_text_factuality | 1467 | 1438 |
| Artificial Analysis Coding Index | 58.6 | 54.5 |
| coding_arena_elo | 1540 | 1460 |
| critpt | 11.3 | 2.9 |
| ERQA | 64.7 | 65.7 |
| frontiermath_tier_4 | 14.6 | 8.3 |
| gdpval | 32.3 | 32 |
| GPQA Diamond | 89.5 | 90.4 |
| HLE | 58.4 | 28.8 |
| IFBench | 75.9 | 75.2 |
| MCP Atlas | 82.2 | 74.1 |
| MMMU-Pro | 80.5 | 78.8 |
| OmniScience Accuracy | 49.6 | 26.4 |
| OmniScience Non-Hallucination | 26.8 | 68 |
| scicode | 51.5 | 40.7 |
| screenspot_pro_no_tools | 84.1 | 68.2 |
| simpleqa_verified | 66.3 | 49.1 |
| SimpleVQA | 0.7 | 0.7 |
| SWE-bench Pro | 55 | 56.6 |
| SWE-bench Verified | 77.4 | 78.8 |
| TauBench V3 - Banking | 19.6 | 20.8 |
| Terminal-Bench 2.0 | 59 | 61.6 |
| Terminal-Bench 2.1 | 62.2 | 61.4 |
| Terminal-Bench Hard | 45.5 | 43.9 |
| τ²-Bench Telecom (AA run) | 91.5 | 97.7 |
| τ³-Bench | 19.6 | 70.7 |
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.