Muse Spark 1.2 vs Qwen3.8 Max
Muse Spark 1.2 
Qwen3.8 Max Verdict
- Qwen3.8 Max (max) leads on quality: 64.2 vs 62.7.
- Muse Spark 1.2 (xhigh) is stronger in human preference, knowledge, maths.
- Qwen3.8 Max (max) is stronger in agents & tools, coding, composite, long context, multimodal, reasoning.
- Muse Spark 1.2 (xhigh) is 1.5× cheaper ($2.00 vs $3.00 per 1M blended).
- Muse Spark 1.2 (xhigh) streams 8.8× faster (325 vs 37 tokens per second).
| Metric | Muse Spark 1.2 (xhigh) | Qwen3.8 Max (max) |
|---|---|---|
| BenchLeader Index | 62.7 | 64.2 |
| Agents & tools score | 55.9 | 58.2 |
| Coding score | 55.0 | 64.6 |
| Composite score | 66.6 | 70.7 |
| Human preference score | 69.1 | 68.0 |
| Knowledge score | 66.9 | 62.7 |
| Long context score | 64.7 | 65.4 |
| Maths score | 66.7 | 65.1 |
| Multimodal score | 65.3 | 66.2 |
| Reasoning score | 69.8 | 72.2 |
| Blended price $/M | $2.00 | $3.00 |
| Output speed | 325 tok/s | 37 tok/s |
| Time to first answer | 18.2 s | 58.9 s |
| Context window | 1.0M | 1M |
| SimpleQA Verified | 60.3% | – |
| Terminal-Bench | – | 27.0% |
| SciCode | 56.4% | 53.2% |
| WeirdML | 60.3% | – |
| APEX-Agents | – | 63.3% |
| ProofBench | – | 58.0% |
| Epoch Capabilities Index | – | 156.4 |
| LMArena Text | 1492 | 1483 |
| LMArena Hard Prompts | 1508 | 1504 |
| LMArena Coding | 1531 | 1524 |
| LMArena WebDev | 1533 | 1672 |
| LMArena Vision | 1305 | 1314 |
| LMArena Agent | -3.7 | 2.3 |
| LiveBench | 78.0% | 78.5% |
| LiveBench Reasoning | 90.0% | 88.2% |
| LiveBench Coding | 77.5% | 72.9% |
| LiveBench Agentic Coding | 57.6% | 64.7% |
| LiveBench Mathematics | 91.2% | 91.3% |
| LiveBench Data Analysis | 76.5% | 78.4% |
| LiveBench Language | 78.6% | 79.7% |
| LiveBench Instruction Following | 74.3% | 74.1% |
| AA Intelligence Index v4.3.2 | 39.6 | 45.4 |
| AA-LCR | 79.0% | 80.3% |
| MMMU-Pro | – | 82.8% |
| AA-Omniscience | 27.2 | 12.0 |
| GPQA Diamond (AA) | 90.4% | 92.8% |
| Humanity's Last Exam (AA) | 45.5% | 43.1% |
| SciCode (AA) | 57.4% | 53.2% |
| LiveCodeBench | – | 87.8% |
| MMLU-Pro | 88.3% | 88.6% |
| IOI | 21.8% | 68.9% |
| LegalBench | 85.3% | 83.6% |
| CorpFin | 70.9% | 65.8% |
| TaxEval | 80.4% | 75.5% |
| Terminal-Bench 2.1 (Vals) | 69.7% | 67.4% |
| SWE-bench (Vals) | 86.6% | 85.6% |
| GPQA Diamond (Vals) | – | 93.7% |
| Vals Index | 49.3 | 48.3 |
| CritPt | 17.7% | 20.0% |
| GDPval-AA v2.1 | 48.9% | 58.6% |
| τ³-Banking (AA) | 34.9% | 51.3% |
| ITBench SRE (AA) | – | 40.3% |
| Analyst Agent (AA) | – | 45.0% |
| APEX-Agents (AA) | – | 42.4% |
| BioMysteryBench | 64.8% | – |
| Code Migration | 29.9% | 24.0% |
| CyberBench | 69.5% | 28.6% |
| Excel Modeling Benchmark | 57.0% | 60.1% |
| Finance Agent v2 | 60.6% | 50.6% |
| Harvey's Legal Agent Benchmark | 25.4% | 10.4% |
| Legal Research Bench | 43.8% | 47.6% |
| MedCode | 49.4% | 40.7% |
| MedScribe | 90.1% | 85.0% |
| MMMU-Pro (Vals) | 86.1% | 88.0% |
| MortgageTax | 65.4% | 64.0% |
| MysteryMechanism | – | 23.9% |
| ProgramBench | – | 0.0% |
| Public Benefits Bench | 68.5% | 67.1% |
| SAGE | 47.7% | 51.3% |
| SkillsBench | 53.0% | 42.0% |
| Tax Agent Bench | 56.9% | 66.0% |
| Terminal-Bench 4.0 (Vals) | 6.1% | 34.3% |
| Terminal-Bench Science | – | 1.4% |
| Vals Multimodal Index | – | 65.4% |
| Vibe Code Bench 1-100 | – | 12.8% |
| Vibe Code Bench v1.1 | 79.1% | 64.7% |
| LMArena Maths | 1481 | 1497 |
| LMArena Creative Writing | 1454 | 1470 |
| LMArena Instruction Following | 1471 | 1474 |
| LMArena Multi-turn | 1506 | 1492 |
| LMArena Longer Queries | 1489 | 1492 |
| DeepSWE v1.1 | 54.9% | – |
| LMCA | 48.4% | – |
| DTBench | 94.7% | – |
| GDP.pdf | 16.0% | – |
| FrontierSWE | 12.0% | – |
| Terminal-Bench 4.0 (AA) | 7.1% | 38.9% |
| Terminal-Bench 2.1 (AA) | 80.2% | 88.8% |
| AutomationBench | – | 56.2% |
| GDP.pdf | – | 22.8% |
| MLCR | – | 20.0% |
| EnterpriseOps-Gym | – | 47.6% |
| AA-Omniscience: accuracy | 45.4% | 31.9% |
| AA-Omniscience: non-hallucination | 66.7% | 71.2% |
| AA-Briefcase v1.1 | – | 1617 |
Data as of 2026-10-11. Best configuration of each model; every score links to its source on the model pages.
Muse Spark 1.2 vs Qwen3.8 Max: questions
- Is Muse Spark 1.2 better than Qwen3.8 Max?
- Qwen3.8 Max (max) leads on quality: 64.2 vs 62.7. The BenchLeader Index combines every independent quality benchmark; Qwen3.8 Max (max) is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is Muse Spark 1.2 better than Qwen3.8 Max for coding?
- Qwen3.8 Max scores higher in coding (65 vs 55 on the category index, where 50 is average).
- Is Muse Spark 1.2 better than Qwen3.8 Max for agentic tasks?
- Qwen3.8 Max scores higher in agentic tasks (58 vs 56 on the category index, where 50 is average).
- Which is cheaper, Muse Spark 1.2 or Qwen3.8 Max?
- Muse Spark 1.2 is cheaper: $2.00 against $3.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Muse Spark 1.2 or Qwen3.8 Max?
- Muse Spark 1.2 streams faster: 325 against 37 output tokens per second.
- Which has the larger context window?
- Muse Spark 1.2 accepts more context: 1.0M against 1M tokens.