GPT-5.6 Sol vs Muse Spark
GPT-5.6 Sol 
Muse Spark Verdict
- GPT-5.6 Sol (high) leads on quality: 68.0 vs 65.4.
- GPT-5.6 Sol (high) is stronger in agents & tools, coding, composite, knowledge, long context, multimodal.
- Muse Spark is stronger in instruction following, reasoning, human preference, maths.
| Metric | GPT-5.6 Sol (high) | Muse Spark |
|---|---|---|
| BenchLeader Index | 68.0 | 65.4 |
| Agents & tools score | 87.6 | 66.3 |
| Coding score | 72.0 | 66.0 |
| Composite score | 82.5 | 68.4 |
| Instruction following score | 67.6 | 79.6 |
| Knowledge score | 73.7 | 64.5 |
| Long context score | 67.4 | 65.5 |
| Multimodal score | 66.5 | 65.9 |
| Reasoning score | 62.9 | 70.0 |
| Human preference score | – | 69.0 |
| Maths score | – | 56.3 |
| Blended price $/M | $8.00 | – |
| Output speed | 59 tok/s | – |
| Time to first answer | 31.4 s | – |
| Context window | 1M | 262k |
| GPQA Diamond | – | 89.8% |
| OTIS Mock AIME | – | 88.9% |
| Humanity's Last Exam | – | 40.6% |
| SciCode | 56.9% | 51.5% |
| WeirdML | 88.8% | – |
| ProofBench | – | 17.0% |
| Epoch Capabilities Index | – | 152.1 |
| LMArena Text | – | 1488 |
| LMArena Hard Prompts | – | 1505 |
| LMArena Coding | – | 1526 |
| LMArena Vision | – | 1306 |
| AA Intelligence Index | 42.5 | 31.3 |
| IFBench | 69.2% | 75.9% |
| AA-LCR | 81.7% | 78.0% |
| MMMU-Pro | 81.8% | 80.5% |
| AA-Omniscience | 20.4 | 7.2 |
| Terminal-Bench Hard | 62.1% | 45.5% |
| GPQA Diamond (AA) | 92.8% | 88.4% |
| Humanity's Last Exam (AA) | 46.0% | 40.7% |
| SciCode (AA) | 57.8% | – |
| τ²-Bench Telecom (AA) | 83.3% | 91.5% |
| AIME (Vals) | – | 96.9% |
| MMLU-Pro | – | 87.3% |
| LegalBench | – | 84.2% |
| CorpFin | – | 65.1% |
| TaxEval | – | 77.7% |
| SWE-bench (Vals) | – | 74.4% |
| GPQA Diamond (Vals) | – | 89.7% |
| SWE-Bench Pro | – | 55.0% |
| MCP Atlas | – | 82.2% |
| MultiChallenge | – | 75.5% |
| PRBench Finance | – | 52.4% |
| PRBench Legal | – | 52.3% |
| EnigmaEval | 37.1% | – |
| MultiNRC | – | 59.0% |
| TutorBench | – | 68.5% |
| ARC-AGI-1 | 97.0% | – |
| ARC-AGI-2 | 85.4% | – |
| ARC-AGI-3 | 2.2% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.