GPT-5.6 Terra vs Muse Wonder
As of 13 Oct 2026, Muse Wonder (max) leads on quality (66.3 vs 63.6 on the BenchLeader Index), but GPT-5.6 Terra (max) costs 1.8× less, so it is the better buy unless you need the extra quality.
GPT-5.6 Terra 
Muse Wonder Key differences
- Muse Wonder (max) leads on quality: 66.3 vs 63.6.
- GPT-5.6 Terra (max) is stronger in agents & tools, coding, instruction following, long context, maths, multimodal.
- Muse Wonder (max) is stronger in composite, knowledge, reasoning.
- GPT-5.6 Terra (max) is 1.8× cheaper ($4.50 vs $8.00 per 1M blended).
- GPT-5.6 Terra (max) streams 1.5× faster (126 vs 85 tokens per second).
Choose GPT-5.6 Terra (max) if
- long context is the priority (67 vs 66 on the category score)
- cost matters: $4.50 against $8.00 per million tokens blended
- you need fast output: 126 against 85 tokens per second
Choose Muse Wonder (max) if
- reasoning is the priority (88 vs 70 on the category score)
- knowledge is the priority (73 vs 57 on the category score)
- the wait for the first word matters: 60.8 s against 131.3 s to first answer
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 | GPT-5.6 Terra (max) | Muse Wonder (max) | Leader |
|---|---|---|---|
| Reasoning | 70.3 | 87.8 | Muse Wonder |
| Coding | 64.7 | – | not comparable |
| Agents & tools | 58.8 | – | not comparable |
| Maths | 69.3 | – | not comparable |
| Knowledge | 57.4 | 73.0 | Muse Wonder |
| Instruction following | 69.8 | – | not comparable |
| Multimodal | 63.9 | – | not comparable |
| Long context | 66.7 | 65.5 | GPT-5.6 Terra |
| Composite | 67.7 | 78.2 | Muse Wonder |
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 | GPT-5.6 Terra (max) | Muse Wonder (max) |
|---|---|---|
| Blended price per 1M tokens | $4.50 | $8.00 |
| Input / output per 1M | $2.00 / $12.00 | $4.00 / $20.00 |
| Price per task: chat reply | $0.0044 | $0.0076 |
| Price per task: summarise a 30-page report | $0.031 | $0.060 |
| Price per task: code edit | $0.030 | $0.054 |
| Price per task: agentic coding session | $0.168 | $0.320 |
| Price per task: structured extraction | $0.0064 | $0.012 |
| Cost to run a full benchmark suite | $1.40 | $4.52 |
| Output speed | 126 tok/s | 85 tok/s |
| Time to first answer | 131.3 s | 60.8 s |
| Time to a complete answer | 135.3 s | 66.7 s |
Every benchmark
| Metric | GPT-5.6 Terra (max) | Muse Wonder (max) |
|---|---|---|
| BenchLeader Index | 63.6 | 66.3 |
| Agents & tools score | 58.8 | – |
| Coding score | 64.7 | – |
| Composite score | 67.7 | 78.2 |
| Instruction following score | 69.8 | – |
| Knowledge score | 57.4 | 73.0 |
| Long context score | 66.7 | 65.5 |
| Maths score | 69.3 | – |
| Multimodal score | 63.9 | – |
| Reasoning score | 70.3 | 87.8 |
| Blended price $/M | $4.50 | $8.00 |
| Output speed | 126 tok/s | 85 tok/s |
| Time to first answer | 131.3 s | 60.8 s |
| Context window | 1.1M | 1.0M |
| GPQA Diamond | 93.3% | – |
| FrontierMath Tiers 1–3 | 86.0% | – |
| FrontierMath Tier 4 | 70.7% | – |
| OTIS Mock AIME | 99.7% | – |
| SimpleQA Verified | 43.2% | – |
| Terminal-Bench | 21.5% | – |
| SciCode | 55.0% | – |
| APEX-Agents | 58.2% | – |
| LiveBench | 77.9% | 81.2% |
| LiveBench Reasoning | 90.6% | 90.6% |
| LiveBench Coding | 78.3% | 80.7% |
| LiveBench Agentic Coding | 55.0% | 62.7% |
| LiveBench Mathematics | 94.9% | 96.5% |
| LiveBench Data Analysis | 79.3% | 81.6% |
| LiveBench Language | 82.9% | 82.9% |
| LiveBench Instruction Following | 64.6% | 73.0% |
| AA Intelligence Index v4.3.2 | 42.1 | 52.4 |
| IFBench | 71.2% | – |
| AA-LCR | 83.0% | 80.7% |
| MMMU-Pro | 80.7% | – |
| AA-Omniscience | 0.1 | 25.5 |
| Terminal-Bench Hard | 57.6% | – |
| GPQA Diamond (AA) | 92.5% | – |
| Humanity's Last Exam (AA) | 42.9% | 50.1% |
| SciCode (AA) | 55.0% | 60.3% |
| τ²-Bench Telecom (AA) | 86.3% | – |
| IOI | 87.6% | – |
| LegalBench | 85.1% | – |
| Terminal-Bench 2.1 (Vals) | 77.5% | – |
| SWE-bench (Vals) | 95.4% | – |
| Vals Index | 53.1 | – |
| ARC-AGI-1 | 96.5% | – |
| ARC-AGI-2 | 83.9% | – |
| ARC-AGI-3 | 0.8% | – |
| CritPt | 30.0% | 28.3% |
| GDPval-AA v2.1 | 47.6% | 63.5% |
| τ³-Banking (AA) | 40.2% | – |
| ITBench SRE (AA) | 51.0% | – |
| APEX-Agents (AA) | 38.9% | – |
| Code Migration | 47.8% | – |
| CyberBench | 72.1% | – |
| Excel Modeling Benchmark | 66.2% | – |
| Finance Agent v2 | 54.4% | – |
| Harvey's Legal Agent Benchmark | 0.8% | – |
| Legal Research Bench | 41.4% | – |
| MMMU-Pro (Vals) | 86.5% | – |
| ProgramBench | 0.5% | – |
| Public Benefits Bench | 62.4% | – |
| SAGE | 47.0% | – |
| SkillsBench | 58.9% | – |
| Tax Agent Bench | 65.2% | – |
| Terminal-Bench 4.0 (Vals) | 22.7% | – |
| Terminal-Bench Science | 10.0% | – |
| Vibe Code Bench 1-100 | 14.8% | – |
| Vibe Code Bench v1.1 | 74.6% | – |
| Chess Puzzles | 54.0% | – |
| Mystery Game Puzzles | 35.0% | – |
| BALROG | 53.2% | – |
| DeepSWE v1.1 | 69.6% | – |
| LMCA | 55.0% | – |
| DTBench | 93.3% | – |
| CursorBench | 41.3% | – |
| ALE-Bench | 1951.4 | – |
| Terminal-Bench 4.0 (AA) | 35.4% | 52.0% |
| Terminal-Bench 2.1 (AA) | 88.0% | – |
| AutomationBench | – | 68.1% |
| GDP.pdf | – | 30.6% |
| AA-Omniscience: accuracy | 46.8% | 41.1% |
| AA-Omniscience: non-hallucination | 12.1% | 73.5% |
| AA-Briefcase v1.1 | – | 1659 |
Data as of 2026-10-13. Best configuration of each model; every score links to its source on the model pages.
GPT-5.6 Terra vs Muse Wonder: questions
- Is GPT-5.6 Terra better than Muse Wonder?
- Muse Wonder (max) leads on quality: 66.3 vs 63.6. The BenchLeader Index combines every independent quality benchmark; Muse Wonder (max) is ahead overall as of 2026-10-13, but check the category scores for your use.
- Which is cheaper, GPT-5.6 Terra or Muse Wonder?
- GPT-5.6 Terra is cheaper: $4.50 against $8.00 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.6 Terra or Muse Wonder?
- GPT-5.6 Terra streams faster: 126 against 85 output tokens per second.
- Which has the larger context window?
- GPT-5.6 Terra accepts more context: 1.1M against 1.0M tokens.