Muse Wonder vs o3
Full profiles: Muse Wonder (Meta) and o3 (OpenAI): every benchmark, price and rank.
Updated · Data from 16 sources
As of 16 Nov 2026, Muse Wonder and o3 are neck and neck on quality (64.4 vs 60.4 on the BenchLeader Index, inside the ±4.4 margin), and o3 is 2.3× cheaper at $3.50 vs $8.00 per 1M tokens blended. Muse Wonder leads in every category we measure both in. o3 streams ~1.5× faster.
Key facts (as of 16 Nov 2026)
- Muse Wonder scores 64.4 on the BenchLeader Index and o3 scores 60.4 (BenchLeader, 16 Nov 2026).
- Muse Wonder costs $8.00 per 1M tokens blended (list price, 16 Nov 2026).
- o3 costs $3.50 per 1M tokens blended (list price, 16 Nov 2026).
- o3 streams ~1.5× faster than Muse Wonder (Artificial Analysis, 7-day median, 16 Nov 2026).
- Of the 5 benchmarks both models have, Muse Wonder leads 5 (BenchLeader, 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 |
|---|---|---|---|
| Agents & tool use | Muse Wonder | 76 vs 70 category score | Medium |
| Maths | Either | 70 vs 71 category score | Either |
| Knowledge | Muse Wonder | 77 vs 56 category score; PRBench Finance 58.7% vs 47.7% | High |
| Chat quality (human preference) | Muse Wonder | 70 vs 61 on Arena votes | High |
| Lowest cost | o3 | $3.50 vs $8.00 per 1M blended; agentic coding session $0.152 vs $0.320 | – |
| Fastest first answer | Muse Wonder | 3.3 s (Muse Wonder) vs 5.2 s (o3) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Muse Wonder and o3 are neck and neck on quality (64.4 vs 60.4, within the ±4.4 margin).
- Muse Wonder is stronger in agents & tools, human preference, knowledge.
- o3 is stronger in maths.
- o3 is 2.3× cheaper ($3.50 vs $8.00 per 1M blended).
- o3 streams 1.9× faster (131 vs 70 tokens per second).
At a glance
| Spec | Muse Wonder | o3 |
|---|---|---|
| Developer | Meta | OpenAI |
| Released | 12 Oct 2026 | 16 Apr 2025 |
| Context window (tokens) | 1,048,576 | 200,000 |
| Compared at | Muse Wonder | o3 |
| Open weights | No | No |
| Price in / out / cached input per 1M | $4.00 / $20.00 / $0.200 | $2.00 / $8.00 / $0.500 |
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 | o3 | Leader |
|---|---|---|---|
| Reasoning | – | 60.9 | not comparable |
| Coding | – | 59.8 | not comparable |
| Agents & tools | 75.6 | 70.3 | Muse Wonder |
| Maths | 70.1 | 70.6 | level |
| Knowledge | 77.1 | 56.2 | Muse Wonder |
| Instruction following | – | 65.2 | not comparable |
| Human preference | 69.6 | 61.0 | Muse Wonder |
| Multimodal | – | 52.8 | not comparable |
| Long context | – | 61.9 | 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: Reasoning, Coding, Instruction following, Long context, Multimodal.
Where sources disagree
Across the 5 benchmarks both have, Muse Wonder leads 5, o3 leads 0. Every overall index that rates both puts Muse Wonder ahead.
| Overall index | Muse Wonder | o3 | Leader |
|---|---|---|---|
| BenchLeader Index | 64.4 | 60.4 | neck and neck |
| Arena Text | 1511† | 1432 | Muse Wonder |
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 | Muse Wonder | o3 |
|---|---|---|
| Blended price per 1M tokens | $8.00 | $3.50 |
| Input / output per 1M | $4.00 / $20.00 | $2.00 / $8.00 |
| Cached input per 1M | $0.200 | $0.500 |
| Price per task: chat reply | $0.0076 | $0.0032 |
| Price per task: summarize a 30-page report | $0.060 | $0.029 |
| Price per task: code edit | $0.054 | $0.024 |
| Price per task: agentic coding session | $0.320 | $0.152 |
| Price per task: structured extraction | $0.012 | $0.0056 |
| Agentic coding session, input mostly cached | $0.149 | $0.085 |
| Output speed | 70 tok/s | 131 tok/s |
| Time to first answer | 3.3 s | 5.2 s |
| Time to a complete answer, last measured 13 Nov 2026 | – | 17.3 s |
Each model is compared at its best-ranked reasoning-effort setting. Muse Wonder scores 64.4 and takes 3 s to its first answer token; Muse Wonder (max) scores 60.8 in 20 s, Muse Wonder (high) scores 56.2 in 13 s. o3 scores 60.4 and takes 5 s to its first answer token; o3 (high) scores 55.1 in 5 s, o3 (medium) scores 53.3 in 5 s. 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
189 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 | Muse Wonder | o3 |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 64.4 | 60.4 |
| Agents & tools score | BenchLeader | 75.6 | 70.3 |
| Human preference score | BenchLeader | 69.6 | 61.0 |
| Knowledge score | BenchLeader | 77.1 | 56.2 |
| Maths score | BenchLeader | 70.1 | 70.6 |
| Coding score | BenchLeader | – | 59.8 |
| Instruction following score | BenchLeader | – | 65.2 |
| Long context score | BenchLeader | – | 61.9 |
| Multimodal score | BenchLeader | – | 52.8 |
| Reasoning score | BenchLeader | – | 60.9 |
| Blended price $/M | BenchLeader | $8.00 | $3.50 |
| Output speed | BenchLeader | 70 tok/s | 131 tok/s |
| Time to first answer | BenchLeader | 3.3 s | 5.2 s |
| Context window | BenchLeader | 1,048,576 tokens | 200,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 81.8%† |
| Humanity's Last Exam | Scale AI / CAIS | – | 20.3%† |
| SimpleBench | SimpleBench | – | 53.1%† |
| Arena Hard Prompts | Arena (LMArena) | 1531† | 1441 |
| LiveBench 2.0 Reasoning | LiveBench | 90.6%† | – |
| LiveBench 1.0 Reasoning | LiveBench | 90.6%† | – |
| GPQA Diamond (AA) | Artificial Analysis | – | 82.7% |
| Humanity's Last Exam (AA) | Artificial Analysis | 50.1%† | 20.1% |
| HLE-Diamond (AA) | Artificial Analysis | 39.0%† | – |
| GPQA Diamond (Vals) | Vals AI | – | 84.1%† |
| Kagi LLM Benchmark | Kagi LLM Benchmark | – | 67.6% |
| GPQA Diamond (HELM) | HELM Capabilities | – | 75.3% |
| ARC-AGI-1 | ARC Prize | – | 60.8%† |
| ARC-AGI-2 | ARC Prize | – | 6.5%† |
| CritPt | Artificial Analysis | 28.3%† | 1.1% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 50.5%† | – |
| MysteryMechanism | Vals AI | 27.5%† | – |
| SciPredict | Scale AI SEAL | – | 17.9% |
| HLE-Diamond (Scale) | Scale AI SEAL | 37.3% | – |
| Arena Expert | Arena (LMArena) | 1525† | 1448 |
| Chess Puzzles | Epoch AI Benchmarking Hub | – | 36.0%† |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 51.0%† | 29.0%† |
| CL-bench | Epoch AI Benchmarking Hub | – | 17.8%† |
| LMCA | Epoch AI Benchmarking Hub | 56.8% | 39.7%† |
| DTBench | Epoch AI Benchmarking Hub | 95.2% | 84.8%† |
Show all 189 benchmarks
| Metric | Publisher | Muse Wonder | o3 |
|---|---|---|---|
| Reasoning | |||
| ForecastBench | ForecastBench | – | 64.3% |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | – | 62.3%† |
| SciCode | SciCode | 60.3%† | – |
| WeirdML | WeirdML | – | 52.4%† |
| GSO-Bench | GSO-Bench | – | 8.8%† |
| Arena Coding | Arena (LMArena) | 1542† | 1460 |
| Arena WebDev | Arena (LMArena) | 1647† | – |
| LiveBench 2.0 Coding | LiveBench | 74.9%† | – |
| LiveBench 1.0 Coding | LiveBench | 80.7%† | – |
| LiveCodeBench | Vals AI | – | 83.9%† |
| IOI | Vals AI | 66.9%† | – |
| Aider Polyglot | Aider polyglot leaderboard | – | 76.9% |
| SWE-bench Verified (bash only) | SWE-bench | – | 58.4% |
| SWE-bench Verified (any scaffold) | SWE-bench | – | 58.4% |
| Code Migration | Vals AI | 62.5%† | – |
| Code Migration v1.1 | Vals AI | 72.4%† | – |
| SWE-bench Multimodal | Vals AI | 45.6%† | – |
| Vibe Code Bench 1-100 | Vals AI | 26.6%† | – |
| Vibe Code Bench v1.1 | Vals AI | 83.1%† | – |
| SWE Atlas v1.1 | Scale AI SEAL | 64.8% | – |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 73.4% | – |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 61.1% | – |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 60.0% | – |
| SWE-Bench Pro V2 | Scale AI SEAL | 96.7%† | – |
| SWE-Bench Pro V2 HARD | Scale AI SEAL | 90.2%† | – |
| MirrorCode | Epoch AI Benchmarking Hub | 45.6%† | – |
| CadEval | Epoch AI Benchmarking Hub | – | 74.0%† |
| CursorBench | Cursor | 50.4%† | – |
| ALE-Bench | Epoch AI Benchmarking Hub | – | 933.5† |
| WeirdML v3 | WeirdML | 15.2%† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 52.0%† | – |
| GameBench 2 Overall | DuelLab GameBench 2 | 64.8 | 34.1 |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 54.9† | 30.4† |
| Design Arena Website | Design Arena | 1378† | 1028 |
| FrontierCode 1.1 (AA) | Artificial Analysis | 46.5%† | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 27.9%† | – |
| LiveCodeBench (AA) | Artificial Analysis | – | 80.8% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 47.0%† | – |
| GDPval | OpenAI | – | 30.8%† |
| APEX-Agents | Mercor | 66.1%† | – |
| Arena Agent | Arena (LMArena) | 1.8† | – |
| LiveBench Long Agentic | LiveBench | 46.2%† | – |
| LiveBench Browser | LiveBench | 60.0%† | – |
| LiveBench Agentic Coding | LiveBench | 62.7%† | – |
| Terminal-Bench Hard | Artificial Analysis | – | 37.1% |
| τ²-Bench Telecom (AA) | Artificial Analysis | – | 80.7% |
| HiL-Bench | Scale AI SEAL | 79.7% | – |
| BFCL Overall | Berkeley Function Calling Leaderboard | – | 63.0% |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 63.5%† | – |
| GDPval-AA v2.1 | Artificial Analysis | 1770† | – |
| CUA-bench | Vals AI | 7.2%† | – |
| CyberBench | Vals AI | 80.4%† | – |
| Finance Agent v2 | Vals AI | 56.6%† | – |
| Finance Agent v2.1 | Vals AI | 65.5%† | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 23.8%† | – |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 17.1%† | – |
| Legal Research Bench | Vals AI | 50.5%† | – |
| Tax Agent Bench | Vals AI | 36.3%† | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 49.5%† | – |
| Terminal-Bench Science | Vals AI | 22.9%† | – |
| Arena Search | Arena (LMArena) | – | 1197 |
| Arena Agent Task Completion | Arena (LMArena) | 5.4† | – |
| DeepResearch Bench | DeepResearch Bench | – | 45.2%† |
| PostTrainBench | Epoch AI Benchmarking Hub | 24.2%† | – |
| METR Time Horizons | METR | – | 63.6% |
| Vending-Bench 2 | Andon Labs | 9432.2 | – |
| AutomationBench | Artificial Analysis | 68.1%† | – |
| GDP.pdf (AA) | Artificial Analysis | 30.6%† | – |
| Harvey LAB-AA v1.1 | Harvey | 10.0%† | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1659† | – |
| GDPx | GDPx by SemiAnalysis | 6.8%† | – |
| Design Arena Agentic | Design Arena | 1177† | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 25.7%† | – |
| MKA (AA) | Artificial Analysis | 47.5%† | – |
| BrowseComp-Plus | BrowseComp-Plus | – | 50.5% |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 82.1%† | 33.3%† |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 70.7%† | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 99.4%† | 83.9%† |
| MATH Level 5 | Epoch AI Benchmarking Hub | – | 97.8%† |
| ProofBench | Vals AI | 84.0% | – |
| LiveBench Mathematics | LiveBench | 96.5%† | – |
| AIME (Vals) | Vals AI | – | 85.3%† |
| MGSM | Vals AI | – | 91.8%† |
| IMO 2025 | MathArena | – | 16.7%† |
| Omni-MATH (HELM) | HELM Capabilities | – | 71.4% |
| MATH 500 | Vals AI | – | 94.6%† |
| Arena Maths | Arena (LMArena) | 1490† | 1448 |
| AIME 2025 (AA) | Artificial Analysis | – | 88.3% |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 49.4%† |
| LiveBench 2.0 Data Analysis | LiveBench | 81.6%† | – |
| LiveBench Research | LiveBench | 93.0%† | – |
| LiveBench 1.0 Data Analysis | LiveBench | 81.6%† | – |
| AA-Omniscience v1.1 | Artificial Analysis | 25.1† | – |
| MMLU-Pro | Vals AI | – | 85.6%† |
| LegalBench | Vals AI | – | 83.8%† |
| CorpFin | Vals AI | – | 59.7%† |
| TaxEval | Vals AI | – | 74.6%† |
| MedQA | Vals AI | – | 96.1%† |
| PRBench Finance | Scale AI SEAL | 58.7% | 47.7% |
| PRBench Legal | Scale AI SEAL | 58.4% | 48.6% |
| MultiNRC | Scale AI SEAL | – | 45.5%† |
| MMLU-Pro (HELM) | HELM Capabilities | – | 85.9% |
| BioMysteryBench | Vals AI | 67.0%† | – |
| Excel Modeling Benchmark | Vals AI | 67.1%† | – |
| Excel Modeling Benchmark v1.1 | Vals AI | 72.5%† | – |
| MedCode | Vals AI | – | 47.3%† |
| MedScribe | Vals AI | – | 76.7%† |
| Public Benefits Bench v1.1 | Vals AI | 70.2%† | – |
| Access to Justice Benchmark | Vals AI | 73.0%† | – |
| Financial Coaching Bench | Vals AI | 73.8%† | – |
| Immigration Bench | Vals AI | 69.6%† | – |
| EBR-bench | Epoch AI Benchmarking Hub | 31.4%† | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 30.8%† | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 38.0%† | – |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 79.1%† | – |
| Instruction following | |||
| LiveBench Language | LiveBench | 82.9%† | – |
| LiveBench Instruction Following | LiveBench | 73.0%† | – |
| IFBench | Artificial Analysis | – | 71.4% |
| MultiChallenge | Scale AI SEAL | – | 56.6%† |
| TutorBench | Scale AI SEAL | – | 52.1%† |
| IFEval (HELM) | HELM Capabilities | – | 86.9% |
| WildBench (HELM) | HELM Capabilities | – | 86.1% |
| Arena Instruction Following | Arena (LMArena) | 1498† | 1403 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1511† | 1432 |
| EQ-Bench 4 | EQ-Bench | 1325 | – |
| EQ-Bench Creative Writing v3 | EQ-Bench | 2007 | 1675 |
| Arena Creative Writing | Arena (LMArena) | 1491† | 1383 |
| Arena Multi-turn | Arena (LMArena) | 1525† | 1420 |
| Arena Longer Queries | Arena (LMArena) | 1507† | 1411 |
| Lech Mazur Creative Writing | Lech Mazur | – | 83.9%† |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | – | 1214 |
| LiveBench Design | LiveBench | 70.9%† | – |
| MMMU-Pro | Artificial Analysis | 83.8%† | 70.1% |
| VISTA | Scale AI SEAL | – | 50.1%† |
| MMMU-Pro (Vals) | Vals AI | – | 80.4%† |
| MortgageTax | Vals AI | – | 65.7%† |
| Humanity's Sixth Sense | Scale AI SEAL | 43.7%† | – |
| VTB | Scale AI SEAL | – | 13.7% |
| GeoBench | GeoBench | – | 60.0%† |
| VPCT | Epoch AI Benchmarking Hub | – | 52.0%† |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 31.7% | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 54.2%† | – |
| VIALS-AA | Handshake AI | 55.3%† | – |
| Code2Video-AA | Artificial Analysis | 997† | – |
| Long context | |||
| Fiction.LiveBench 120k | Fiction.live | – | 100.0%† |
| AA-LCR v1.1 | Artificial Analysis | 80.7%† | 74.7% |
| MLCR | Artificial Analysis | 22.2%† | – |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 161.7 | 147.1 |
| LiveBench 2.0 | LiveBench | 70.7%† | – |
| LiveBench 1.0 | LiveBench | 81.2%† | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 43.6† | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 52.4† | 20.2 |
| AA Progress Index v1 | Artificial Analysis | 335.8† | 258.3 |
| Vals Index | Vals AI | 64.9† | – |
| HELM Capabilities mean | HELM Capabilities | – | 81.1% |
| Vals Finance Index | Vals AI | 61.2%† | – |
| Vals Science Index | Vals AI | 39.1%† | – |
| Vals Router Index | Vals AI | 64.9% | – |
| Vals Legal Index | Vals AI | 75.3%† | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | – | 16.0% |
| MASK | Scale AI SEAL | – | 84.5%† |
| PropensityBench | Scale AI SEAL | – | 10.5% |
| DistressBench | Scale AI SEAL | 85.9% | – |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 5.6 |
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
- Muse Wonder: BenchLeader Index 63.0 → 64.4.
- o3: BenchLeader Index 61.2 → 60.4.
- o3: 1 score revision.
- Muse Wonder: BenchLeader Index 63.5 → 63.0.
- Muse Wonder: 1 new result (Blueprint-Bench 2).
- Muse Wonder: BenchLeader Index 64.3 → 63.5.
- Muse Wonder: 5 score revisions.
- o3: 1 new result (Arena Search).
- Muse Wonder: 1 score revision.
- o3: 1 score revision.
Related comparisons
- Muse Wonder vs Claude Fable 5.2
- Muse Wonder vs Gemini 4 Carbon
- Muse Wonder vs Claude Opus 5.5
- Muse Wonder vs GPT-6 Astra
- o3 vs Claude Opus 5.5
- Muse Spark 1.3 vs o3
- Muse Wonder vs Claude Fable 5.1
- Muse Wonder vs GPT-6.1 Sol
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Muse Wonder vs o3: questions
- Is Muse Wonder better than o3?
- Muse Wonder and o3 are neck and neck on quality (64.4 vs 60.4, within the ±4.4 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
- Muse Wonder vs o3: which is cheaper, and will that change?
- o3 is cheaper today: $3.50 against $8.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Which is faster, Muse Wonder or o3?
- o3 streams ~1.5× faster (131 vs 70 output tokens per second, Artificial Analysis 7-day median).
- Is Muse Wonder or o3 better for agents and tool use?
- Muse Wonder scores higher in agentic tasks (76 vs 70 on the category index, where 50 is average).
- Which is better for long documents?
- Muse Wonder accepts more context: 1,048,576 against 200,000 tokens for o3.
More questions (3)
- Which has lower latency?
- Muse Wonder answers first: 3.3 s against 5.2 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on Muse Wonder vs o3?
- Across the 5 benchmarks both have, Muse Wonder leads 5, o3 leads 0. Every overall index that rates both puts Muse Wonder ahead.
- When was Muse Wonder released?
- Muse Wonder was released on 12 Oct 2026.