Muse Wonder vs Muse Wonder Full
As of 18 Oct 2026, Muse Wonder Full (max) is the better pick for most workloads: it leads on quality (68.6 vs 66.8 on the BenchLeader Index).
Muse Wonder Key differences
- Muse Wonder Full (max) leads on quality: 68.6 vs 66.8.
- Muse Wonder (max) is stronger in agents & tools, coding, human preference, knowledge, maths.
- Muse Wonder Full (max) is stronger in composite, long context, reasoning.
Choose Muse Wonder (max) if
- coding is the priority (66 vs 64 on the category score)
- knowledge is the priority (73 vs 71 on the category score)
Choose Muse Wonder Full (max) if
- reasoning is the priority (95 vs 82 on the category score)
- long context is the priority (67 vs 65 on the category score)
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Muse Wonder (max) | Muse Wonder Full (max) | Leader |
|---|---|---|---|
| Reasoning | 81.8 | 95.0 | Muse Wonder Full |
| Coding | 66.3 | 64.0 | Muse Wonder |
| Agents & tools | 57.7 | – | not comparable |
| Maths | 67.8 | – | not comparable |
| Knowledge | 72.9 | 71.3 | Muse Wonder |
| Human preference | 71.0 | – | not comparable |
| Long context | 65.4 | 67.0 | Muse Wonder Full |
| Composite | 78.1 | 92.2 | Muse Wonder Full |
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | Muse Wonder (max) | Muse Wonder Full (max) |
|---|---|---|
| Blended price per 1M tokens | $8.00 | – |
| Input / output per 1M | $4.00 / $20.00 | – / – |
| Price per task: chat reply | $0.0076 | – |
| Price per task: summarise a 30-page report | $0.060 | – |
| Price per task: code edit | $0.054 | – |
| Price per task: agentic coding session | $0.320 | – |
| Price per task: structured extraction | $0.012 | – |
| Cost to run a full benchmark suite | $4.52 | – |
| Output speed | 176 tok/s | – |
| Time to first answer | 47.3 s | – |
| Time to a complete answer | 50.2 s | – |
Every benchmark
| Metric | Muse Wonder (max) | Muse Wonder Full (max) |
|---|---|---|
| BenchLeader Index | 66.8 | 68.6 |
| Agents & tools score | 57.7 | – |
| Coding score | 66.3 | 64.0 |
| Composite score | 78.1 | 92.2 |
| Human preference score | 71.0 | – |
| Knowledge score | 72.9 | 71.3 |
| Long context score | 65.4 | 67.0 |
| Maths score | 67.8 | – |
| Reasoning score | 81.8 | 95.0 |
| Blended price $/M | $8.00 | – |
| Output speed | 176 tok/s | – |
| Time to first answer | 47.3 s | – |
| Context window | 1.0M | 1M |
| FrontierMath Tiers 1–3 | 82.1% | – |
| FrontierMath Tier 4 | 70.7% | – |
| OTIS Mock AIME | 99.4% | – |
| SciCode | 60.3% | – |
| Arena Text | 1509 | – |
| Arena Hard Prompts | 1527 | – |
| Arena Coding | 1548 | – |
| Arena WebDev | 1650 | – |
| Arena Agent | 4.9 | – |
| LiveBench | 81.2% | – |
| LiveBench Reasoning | 90.6% | – |
| LiveBench Coding | 80.7% | – |
| LiveBench Agentic Coding | 62.7% | – |
| LiveBench Mathematics | 96.5% | – |
| LiveBench Data Analysis | 81.6% | – |
| LiveBench Language | 82.9% | – |
| LiveBench Instruction Following | 73.0% | – |
| AA Intelligence Index v4.3.2 | 52.4 | 55.8 |
| AA-LCR | 80.7% | 83.7% |
| AA-Omniscience | 25.5 | 30.1 |
| Humanity's Last Exam (AA) | 50.1% | 53.6% |
| SciCode (AA) | 60.3% | 62.6% |
| IOI | 66.9% | 79.8% |
| LegalBench | – | 87.4% |
| Vals Index | 61.1 | 64.6 |
| CritPt | 28.3% | 32.9% |
| GDPval-AA v2.1 | 63.5% | 67.8% |
| BioMysteryBench | 67.0% | 73.3% |
| Code Migration | 62.5% | 61.0% |
| CUA-bench | 7.2% | – |
| CyberBench | 80.4% | 73.6% |
| Excel Modeling Benchmark | 67.1% | 71.9% |
| Finance Agent v2 | 56.6% | 59.1% |
| Harvey's Legal Agent Benchmark | 17.1% | 21.7% |
| Legal Research Bench | 50.5% | 53.4% |
| MedCode | – | 49.9% |
| MedScribe | – | 89.3% |
| MysteryMechanism | 27.5% | 35.1% |
| ProgramBench | – | 0.5% |
| Public Benefits Bench | 70.2% | 73.8% |
| SREBench | – | 11.8% |
| Tax Agent Bench | 36.3% | 43.0% |
| Terminal-Bench 4.0 (Vals) | 49.5% | 58.1% |
| Terminal-Bench Science | 22.9% | 37.1% |
| Vibe Code Bench 1-100 | 26.6% | 34.0% |
| Vibe Code Bench v1.1 | 83.1% | 86.0% |
| Arena Creative Writing | 1485 | – |
| Arena Instruction Following | 1492 | – |
| Arena Multi-turn | 1530 | – |
| Arena Longer Queries | 1500 | – |
| EBR-bench | 31.4% | – |
| Mystery Game Puzzles | 51.0% | – |
| GDP.pdf | 30.8% | – |
| Terminal-Bench 4.0 (AA) | 52.0% | 57.1% |
| AutomationBench | 68.1% | 71.3% |
| GDP.pdf | 30.6% | 32.6% |
| Harvey LAB-AA v1.1 | 10.0% | 13.1% |
| AA-Omniscience: accuracy | 41.1% | 43.2% |
| AA-Omniscience: non-hallucination | 73.5% | 77.0% |
| AA-Briefcase v1.1 | 1659 | 1729 |
Data as of 2026-10-18. Best configuration of each model; every score links to its source on the model pages.
Muse Wonder vs Muse Wonder Full: questions
- Is Muse Wonder better than Muse Wonder Full?
- Muse Wonder Full (max) leads on quality: 68.6 vs 66.8. The BenchLeader Index combines every independent quality benchmark; Muse Wonder Full (max) is ahead overall as of 2026-10-18, but check the category scores for your use.
- Is Muse Wonder better than Muse Wonder Full for coding?
- Muse Wonder scores higher in coding (66 vs 64 on the category index, where 50 is average).
- Which has the larger context window?
- Muse Wonder accepts more context: 1.0M against 1M tokens.