Claude Opus 4.8 vs Muse Spark
Claude Opus 4.8 
Muse Spark Verdict
- Claude Opus 4.8 and Muse Spark are level on quality (64.7 vs 65.4).
- Claude Opus 4.8 is stronger in composite, knowledge.
- Muse Spark is stronger in agents & tools, coding, human preference, instruction following, long context, multimodal, reasoning, maths.
| Metric | Claude Opus 4.8 | Muse Spark |
|---|---|---|
| BenchLeader Index | 64.7 | 65.4 |
| Agents & tools score | 65.0 | 66.3 |
| Coding score | 65.7 | 66.0 |
| Composite score | 81.9 | 68.4 |
| Human preference score | 65.6 | 69.0 |
| Instruction following score | 61.6 | 79.6 |
| Knowledge score | 66.3 | 64.5 |
| Long context score | 65.3 | 65.5 |
| Multimodal score | 64.4 | 65.9 |
| Reasoning score | 64.2 | 70.0 |
| Maths score | – | 56.3 |
| Blended price $/M | $10.00 | – |
| Output speed | 58 tok/s | – |
| Time to first answer | 29.5 s | – |
| Context window | 1M | 262k |
| GPQA Diamond | – | 89.8% |
| OTIS Mock AIME | – | 88.9% |
| Humanity's Last Exam | – | 40.6% |
| SimpleBench | 64.8% | – |
| SciCode | – | 51.5% |
| Remote Labor Index | 8.3% | – |
| FrontierCode | 46.5% | – |
| ProofBench | – | 17.0% |
| GSO-Bench | 47.1% | – |
| Epoch Capabilities Index | 158.3 | 152.1 |
| LMArena Text | 1473 | 1488 |
| LMArena Hard Prompts | 1503 | 1505 |
| LMArena Coding | 1527 | 1526 |
| LMArena WebDev | 1540 | – |
| LMArena Vision | 1289 | 1306 |
| AA Intelligence Index | 42.0 | 31.3 |
| IFBench | 62.2% | 75.9% |
| AA-LCR | 77.7% | 78.0% |
| MMMU-Pro | – | 80.5% |
| AA-Omniscience | 28.8 | 7.2 |
| Terminal-Bench Hard | 58.3% | 45.5% |
| GPQA Diamond (AA) | 92.0% | 88.4% |
| Humanity's Last Exam (AA) | 48.7% | 40.7% |
| SciCode (AA) | 54.4% | – |
| τ²-Bench Telecom (AA) | 94.4% | 91.5% |
| AIME (Vals) | – | 96.9% |
| LiveCodeBench | 87.8% | – |
| MMLU-Pro | 89.6% | 87.3% |
| LegalBench | 83.6% | 84.2% |
| CorpFin | 66.7% | 65.1% |
| TaxEval | 75.6% | 77.7% |
| Terminal-Bench 2.1 (Vals) | 71.9% | – |
| SWE-bench (Vals) | 88.6% | 74.4% |
| GPQA Diamond (Vals) | 92.4% | 89.7% |
| Vals Index | 60.9 | – |
| SWE-Bench Pro | – | 55.0% |
| MCP Atlas | – | 82.2% |
| MultiChallenge | – | 75.5% |
| PRBench Finance | – | 52.4% |
| PRBench Legal | – | 52.3% |
| MultiNRC | – | 59.0% |
| HiL-Bench | 35.3% | – |
| TutorBench | – | 68.5% |
| EQ-Bench 4 | 1281 | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.