GPT-6 Luna vs Muse Wonder Full
Full profiles: GPT-6 Luna (OpenAI) and Muse Wonder Full (Meta): every benchmark, price and rank.
Updated · Data from 12 sources
As of 16 Nov 2026, Muse Wonder Full is ahead of GPT-6 Luna on quality (61.7 vs 55.1 on the BenchLeader Index). Muse Wonder Full leads in every category we measure both in.
Key facts (as of 16 Nov 2026)
- GPT-6 Luna scores 55.1 on the BenchLeader Index and Muse Wonder Full scores 61.7 (BenchLeader, 16 Nov 2026).
- GPT-6 Luna costs $0.200 per 1M tokens blended (list price, 16 Nov 2026).
- Of the 53 benchmarks both models have, Muse Wonder Full leads 50 (BenchLeader, 16 Nov 2026).
- On Arena WebDev, GPT-6 Luna is rated 1577 and Muse Wonder Full 1724 (LMArena, as published by 16 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Muse Wonder Full | 70 vs 57 category score; Arena WebDev 1724 vs 1577 | High |
| Agents & tool use | Muse Wonder Full | 64 vs 49 category score; Arena Agent 9 vs 1.6 | High |
| Reasoning | Muse Wonder Full | 66 vs 56 category score; HLE-Diamond (AA) 48.5% vs 22.2% | High |
| Knowledge | Muse Wonder Full | 62 vs 45 category score; AA-Omniscience v1.1 30.6 vs -8.3 | High |
| Long documents | Either | 66 vs 67 category score; AA-LCR v1.1 83.3% vs 83.7% | Either |
| Chat quality (human preference) | – | Only GPT-6 Luna has an Arena rating (62 category score) | Not comparable |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Muse Wonder Full (max) leads on quality: 61.7 vs 55.1.
- Muse Wonder Full (max) is stronger in agents & tools, coding, knowledge, long context, reasoning.
At a glance
| Spec | GPT-6 Luna | Muse Wonder Full |
|---|---|---|
| Developer | OpenAI | Meta |
| Released | 22 Sep 2026 | 12 Oct 2026 |
| Context window (tokens) | 1,050,000 | 1,000,000 |
| Compared at | GPT-6 Luna (max) | Muse Wonder Full (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $0.100 / $0.500 / $0.010 | – |
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-6 Luna (max) | Muse Wonder Full (max) | Leader |
|---|---|---|---|
| Reasoning | 55.7 | 65.8 | Muse Wonder Full |
| Coding | 57.3 | 69.9 | Muse Wonder Full |
| Agents & tools | 48.5 | 63.6 | Muse Wonder Full |
| Maths | 63.6 | – | not comparable |
| Knowledge | 44.8 | 62.3 | Muse Wonder Full |
| Human preference | 62.2 | – | not comparable |
| Multimodal | 57.6 | – | not comparable |
| Long context | 66.4 | 66.6 | level |
| Composite | 44.5 | – | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Maths, Instruction following, Multimodal, Human preference.
Where sources disagree
Across the 53 benchmarks both have, GPT-6 Luna leads 2, Muse Wonder Full leads 50 and 1 is a tie. Every overall index that rates both puts Muse Wonder Full ahead.
| Overall index | GPT-6 Luna (max) | Muse Wonder Full (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 55.1 | 61.7 | Muse Wonder Full |
| AA Intelligence Index v5.0.1 | 27.9 | 50.1 | Muse Wonder Full |
| Vals Index | 52.3 | 66.8 | Muse Wonder Full |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
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. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | GPT-6 Luna (max) | Muse Wonder Full (max) |
|---|---|---|
| Blended price per 1M tokens | $0.200 | – |
| Input / output per 1M | $0.100 / $0.500 | – / – |
| Cached input per 1M | $0.010 | – |
| Price per task: chat reply | $0.0002 | – |
| Price per task: summarize a 30-page report | $0.0015 | – |
| Price per task: code edit | $0.0014 | – |
| Price per task: agentic coding session | $0.0080 | – |
| Price per task: structured extraction | $0.0003 | – |
| Agentic coding session, input mostly cached | $0.0040 | – |
| Cost per benchmark task (AA Index) | $0.152 | – |
| Output tokens per task (AA Index) | 75k | 88k |
| Of which reasoning | 58k | 55k |
| Time per task (AA Index) | 12 min | – |
| Output speed | 103 tok/s | – |
| Time to first answer | 73.9 s | – |
| Time to a complete answer, last measured 13 Nov 2026 | 171.6 s | – |
Each model is compared at its best-ranked reasoning-effort setting. GPT-6 Luna (max) scores 55.1 and takes 74 s to its first answer token; GPT-6 Luna (xhigh) scores 52.0 in 17 s, GPT-6 Luna (high) scores 51.1 in 10 s. Muse Wonder Full (max) scores 61.7. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
145 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | GPT-6 Luna (max) | Muse Wonder Full (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 55.1 | 61.7 |
| Agents & tools score | BenchLeader | 48.5 | 63.6 |
| Coding score | BenchLeader | 57.3 | 69.9 |
| Composite score | BenchLeader | 44.5 | – |
| Human preference score | BenchLeader | 62.2 | – |
| Knowledge score | BenchLeader | 44.8 | 62.3 |
| Long context score | BenchLeader | 66.4 | 66.6 |
| Maths score | BenchLeader | 63.6 | – |
| Multimodal score | BenchLeader | 57.6 | – |
| Reasoning score | BenchLeader | 55.7 | 65.8 |
| Blended price $/M | BenchLeader | $0.200 | – |
| Output speed | BenchLeader | 103 tok/s | – |
| Time to first answer | BenchLeader | 73.9 s | – |
| Context window | BenchLeader | 1,050,000 tokens | 1,000,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 90.5% | – |
| Arena Hard Prompts | Arena (LMArena) | 1464 | – |
| LiveBench 2.0 Reasoning | LiveBench | 81.8% | – |
| LiveBench 1.0 Reasoning | LiveBench | 81.8% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 38.5% | 53.6% |
| HLE-Diamond (AA) | Artificial Analysis | 22.2% | 48.5% |
| ARC-AGI-1 | ARC Prize | 86.7% | – |
| ARC-AGI-2 | ARC Prize | 59.3% | – |
| ARC-AGI-3 | ARC Prize | 0.6% | – |
| CritPt | Artificial Analysis | 19.4% | 32.9% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 42.5% | 56.5% |
| MysteryMechanism | Vals AI | 19.4% | 35.1% |
| Arena Expert | Arena (LMArena) | 1487 | – |
| Chess Puzzles | Epoch AI Benchmarking Hub | 36.0% | – |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 7.0% | – |
| LMCA | Epoch AI Benchmarking Hub | 44.5% | – |
| DTBench | Epoch AI Benchmarking Hub | 90.1% | – |
| Coding | |||
| SciCode | SciCode | 54.6% | – |
| FrontierCode | Cognition | 42.4% | – |
| Arena Coding | Arena (LMArena) | 1501 | – |
| Arena WebDev | Arena (LMArena) | 1577 | 1724 |
| LiveBench 2.0 Coding | LiveBench | 70.8% | – |
| LiveBench 1.0 Coding | LiveBench | 79.0% | – |
| SciCode (AA) | Artificial Analysis | 54.6% | 62.6% |
| IOI | Vals AI | 55.6% | 79.8% |
Show all 145 benchmarks
| Metric | Publisher | GPT-6 Luna (max) | Muse Wonder Full (max) |
|---|---|---|---|
| Coding | |||
| Code Migration | Vals AI | 42.5% | 61.0% |
| Code Migration v1.1 | Vals AI | 46.7% | 62.8% |
| ProgramBench | Vals AI | 0.5% | 0.5% |
| SWE-bench Multimodal | Vals AI | 37.5% | 47.3% |
| Vibe Code Bench 1-100 | Vals AI | – | 34.0% |
| Vibe Code Bench v1.1 | Vals AI | 81.7% | 86.0% |
| SWE Atlas v1.1 | Scale AI SEAL | 50.8%† | – |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 50.8%† | – |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 54.4%† | – |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 47.1%† | – |
| MirrorCode | Epoch AI Benchmarking Hub | 26.7% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 1576.9† | – |
| WeirdML v3 | WeirdML | 7.9%† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 12.6% | 57.1% |
| Startrise Frontend | Startrise | 69.3%† | – |
| GameBench 2 Overall | DuelLab GameBench 2 | 47.5† | – |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 55.9† | – |
| Design Arena Website | Design Arena | 1268† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 41.9% | 44.9% |
| FrontierCode 2.0 (AA) | Artificial Analysis | 25.2% | 37.8% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 15.2% | – |
| APEX-Agents | Mercor | 44.3% | – |
| Arena Agent | Arena (LMArena) | 1.6 | 9 |
| LiveBench Long Agentic | LiveBench | 24.7% | – |
| LiveBench Browser | LiveBench | 70.3% | – |
| LiveBench Agentic Coding | LiveBench | 51.2% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 73.0% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 46.6% | 67.8% |
| GDPval-AA v2.1 | Artificial Analysis | 1443 | 1857 |
| Analyst Agent (AA) | Artificial Analysis | 43.8% | – |
| CyberBench | Vals AI | 76.3% | 73.6% |
| Finance Agent v2 | Vals AI | 49.9% | 59.1% |
| Finance Agent v2.1 | Vals AI | 56.8% | 67.6% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 4.2% | 23.3% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 2.9% | 21.7% |
| Legal Research Bench | Vals AI | 30.3% | 53.4% |
| Robo-bench | Vals AI | – | 10.4% |
| SREBench | Vals AI | 2.7% | 11.8% |
| Tax Agent Bench | Vals AI | 19.9% | 43.0% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 13.6% | 58.1% |
| Terminal-Bench Science | Vals AI | 4.3% | 37.1% |
| Arena Search | Arena (LMArena) | 1211 | – |
| Arena Agent Task Completion | Arena (LMArena) | -0.4 | 11.1 |
| SpireClimb | Epoch AI Benchmarking Hub | 0.0% | – |
| AutomationBench | Artificial Analysis | 53.2% | 71.3% |
| GDP.pdf (AA) | Artificial Analysis | 22.8% | 32.6% |
| Harvey LAB-AA v1.1 | Harvey | 3.3% | 13.1% |
| AA-Briefcase v1.1 | Artificial Analysis | 1337 | 1729 |
| GDPx | GDPx by SemiAnalysis | 5.1% | 9.0% |
| Design Arena Agentic | Design Arena | 1152† | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 8.6% | 37.1% |
| CWE-Bench (AA) | Artificial Analysis | 56.7% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 23.5% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 77.9% | – |
| MKA (AA) | Artificial Analysis | 36.8% | – |
| FACTS Search | 86.8%† | – | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 79.0% | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 56.1% | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 98.9% | – |
| ProofBench | Vals AI | 64.0%† | – |
| LiveBench Mathematics | LiveBench | 89.1% | – |
| Arena Maths | Arena (LMArena) | 1449 | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 41.4% | – |
| LiveBench 2.0 Data Analysis | LiveBench | 73.4% | – |
| LiveBench Research | LiveBench | 48.5% | – |
| LiveBench 1.0 Data Analysis | LiveBench | 73.4% | – |
| AA-Omniscience v1.1 | Artificial Analysis | -8.3 | 30.6 |
| LegalBench | Vals AI | – | 87.4% |
| PRBench Legal | Scale AI SEAL | 42.9%† | – |
| BioMysteryBench | Vals AI | 61.5% | 73.3% |
| ClaimBench | Vals AI | 99.2% | – |
| Excel Modeling Benchmark | Vals AI | 68.5% | 71.9% |
| Excel Modeling Benchmark v1.1 | Vals AI | 70.0% | 75.0% |
| MedCode | Vals AI | 44.7% | 49.9% |
| MedScribe | Vals AI | 83.7% | 89.3% |
| Public Benefits Bench v1.1 | Vals AI | 57.6% | 73.8% |
| Access to Justice Benchmark | Vals AI | – | 74.7% |
| Financial Coaching Bench | Vals AI | 65.2% | 75.0% |
| Immigration Bench | Vals AI | 62.5% | 71.2% |
| GDP.pdf | Epoch AI Benchmarking Hub | 23.0% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 39.6% | 41.0% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 20.6% | 82.4% |
| HealthBench Professional (AA) | OpenAI | 61.9% | – |
| FACTS Parametric | 57.0%† | – | |
| Instruction following | |||
| LiveBench Language | LiveBench | 73.8% | – |
| LiveBench Instruction Following | LiveBench | 55.9% | – |
| Arena Instruction Following | Arena (LMArena) | 1436 | – |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1441 | – |
| Arena Creative Writing | Arena (LMArena) | 1408 | – |
| Arena Multi-turn | Arena (LMArena) | 1441 | – |
| Arena Longer Queries | Arena (LMArena) | 1449 | – |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1211 | – |
| LiveBench Design | LiveBench | 76.1% | – |
| MMMU-Pro | Artificial Analysis | 79.7% | – |
| SAGE | Vals AI | 83.4% | – |
| Humanity's Sixth Sense | Scale AI SEAL | 21.0% | – |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 31.2%† | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 44.2% | – |
| VIALS-AA | Handshake AI | 27.3% | – |
| Code2Video-AA | Artificial Analysis | 1030 | – |
| AA-3D | Artificial Analysis | 1000 | – |
| FACTS Multimodal | 38.7%† | – | |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 83.3% | 83.7% |
| MLCR | Artificial Analysis | 26.1% | 22.2% |
| FACTS Grounding | 76.7%† | – | |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 156.4† | – |
| LiveBench 2.0 | LiveBench | 60.4% | – |
| LiveBench 1.0 | LiveBench | 72.0% | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 27.9 | 50.1 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 38.1 | 55.8 |
| AA Progress Index v1 | Artificial Analysis | 317.1 | 343.2 |
| Vals Index | Vals AI | 52.3 | 66.8 |
| Vals Finance Index | Vals AI | 53.0% | 64.5% |
| Vals RSI Index | Vals AI | – | 24.4% |
| Vals Science Index | Vals AI | 28.4% | 48.5% |
| Vals Router Index | Vals AI | 52.3%† | 66.8%† |
| Vals Legal Index | Vals AI | 62.2% | 77.9% |
| AA Cyber Index | Artificial Analysis | 52.7 | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 10.8%† | – |
| DistressBench | Scale AI SEAL | 66.7%† | – |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- GPT-6 Luna (max): BenchLeader Index 56.6 → 55.1.
- GPT-6 Luna (max): 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
- Muse Wonder Full (max): BenchLeader Index 62.6 → 61.7.
- GPT-6 Luna (max): 1 score revision.
- GPT-6 Luna (max): 1 score revision.
- GPT-6 Luna (max): 1 score revision.
- Muse Wonder Full (max): 1 score revision.
- Muse Wonder Full (max): 1 score revision.
- GPT-6 Luna (max): 1 score revision.
- Muse Wonder Full (max): 1 new result (Arena Agent).
Related comparisons
- GPT-6 Luna vs GPT-6 Astra
- GPT-6 Luna vs GPT-6.1 Sol
- GPT-6 Luna vs Gemini 4 Argon
- Muse Wonder Full vs Claude Fable 5.2
- Muse Wonder Full vs Gemini 4 Carbon
- Muse Wonder Full vs Claude Opus 5.5
- Muse Wonder Full vs GPT-6 Astra
- GPT-6 Luna vs GPT-5.5
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
GPT-6 Luna vs Muse Wonder Full: questions
- Is Muse Wonder Full better than GPT-6 Luna?
- Muse Wonder Full (max) leads on quality: 61.7 vs 55.1. The BenchLeader Index combines every independent quality benchmark; Muse Wonder Full (max) is ahead overall as of 16 Nov 2026, but check the category scores for your use.
- Is GPT-6 Luna or Muse Wonder Full better for coding?
- Muse Wonder Full scores higher in coding (70 vs 57 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1577 for GPT-6 Luna against 1724 for Muse Wonder Full.
- Is GPT-6 Luna or Muse Wonder Full better for agents and tool use?
- Muse Wonder Full scores higher in agentic tasks (64 vs 49 on the category index, where 50 is average). On Arena Agent, the top shared agents & tools benchmark, it is 1.6 for GPT-6 Luna against 9 for Muse Wonder Full.
- Which is better for long documents?
- They are level on long-context benchmarks (66 vs 67). Both take about 1.1M tokens: GPT-6 Luna 1,050,000, Muse Wonder Full 1,000,000.
- Do independent benchmarks agree on GPT-6 Luna vs Muse Wonder Full?
- Across the 53 benchmarks both have, GPT-6 Luna leads 2, Muse Wonder Full leads 50 and 1 is a tie. Every overall index that rates both puts Muse Wonder Full ahead.
More questions (1)
- When was Muse Wonder Full released?
- Muse Wonder Full was released on 12 Oct 2026.