Gemini 4 Carbon vs Muse Wonder
Full profiles: Gemini 4 Carbon (Google) and Muse Wonder (Meta): every benchmark, price and rank.
Updated · Data from 9 sources
As of 29 Oct 2026, Gemini 4 Carbon and Muse Wonder are neck and neck on quality (65.1 vs 64.0 on the BenchLeader Index, inside the ±7.9 margin), and they cost about the same at $8.00 vs $8.00 per 1M tokens blended. They are level in every category we measure both in.
Key facts (as of 29 Oct 2026)
- Gemini 4 Carbon scores 65.1 on the BenchLeader Index and Muse Wonder scores 64.0 (BenchLeader, 29 Oct 2026).
- Gemini 4 Carbon costs $8.00 per 1M tokens blended (list price, 29 Oct 2026).
- Muse Wonder costs $8.00 per 1M tokens blended (list price, 29 Oct 2026).
Gemini 4 Carbon benchmarks 
Muse Wonder benchmarks 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 |
|---|---|---|---|
| Knowledge | Either | 76 vs 76 category score | Either |
| Chat quality (human preference) | – | Only Muse Wonder has an Arena rating (70 category score) | Not comparable |
| Lowest cost | Either | $8.00 vs $8.00 per 1M blended; agentic coding session $0.320 vs $0.320 | – |
| Fastest first answer | Muse Wonder | 3.6 s (Muse Wonder) vs 103.1 s (Gemini 4 Carbon (high)) 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
- Gemini 4 Carbon (high) and Muse Wonder are neck and neck on quality (65.1 vs 64.0, within the ±7.9 margin).
- Gemini 4 Carbon (high) is stronger in coding, long context, reasoning.
- Muse Wonder is stronger in knowledge, agents & tools, composite, human preference, maths.
- They cost about the same ($8.00 per 1M blended).
At a glance
| Spec | Gemini 4 Carbon | Muse Wonder |
|---|---|---|
| Developer | Meta | |
| Released | 28 Oct 2026 | 12 Oct 2026 |
| Context window (tokens) | 1,000,000 | 1,048,576 |
| Compared at | Gemini 4 Carbon (high) | Muse Wonder |
| Open weights | No | No |
| Price in / out / cached input per 1M | $4.00 / $20.00 / – | $4.00 / $20.00 / $0.200 |
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 | Gemini 4 Carbon (high) | Muse Wonder | Leader |
|---|---|---|---|
| Reasoning | 78.9 | – | not comparable |
| Coding | 71.7 | – | not comparable |
| Agents & tools | – | 78.7 | not comparable |
| Maths | – | 70.3 | not comparable |
| Knowledge | 75.8 | 76.4 | level |
| Human preference | – | 69.6 | not comparable |
| Long context | 64.5 | – | not comparable |
| Composite | – | 60.2 | not comparable |
Where sources disagree
Across the 0 benchmarks both have, Gemini 4 Carbon leads 0, Muse Wonder leads 0. Every overall index that rates both puts Gemini 4 Carbon ahead.
| Overall index | Gemini 4 Carbon (high) | Muse Wonder | Leader |
|---|---|---|---|
| BenchLeader Index | 65.1 | 64.0 | neck and neck |
| AA Intelligence Index v5.0 | 59.4 | 43.3† | Gemini 4 Carbon |
| Vals Index | 70.4 | 64.9† | Gemini 4 Carbon |
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; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | Gemini 4 Carbon (high) | Muse Wonder |
|---|---|---|
| Blended price per 1M tokens | $8.00 | $8.00 |
| Input / output per 1M | $4.00 / $20.00 | $4.00 / $20.00 |
| Cached input per 1M | – | $0.200 |
| Price per task: chat reply | $0.0076 | $0.0076 |
| Price per task: summarize a 30-page report | $0.060 | $0.060 |
| Price per task: code edit | $0.054 | $0.054 |
| Price per task: agentic coding session | $0.320 | $0.320 |
| Price per task: structured extraction | $0.012 | $0.012 |
| Cost to run a full benchmark suite | $7.76 | – |
| Output speed | 100 tok/s | 81 tok/s |
| Time to first answer | 103.1 s | 3.6 s |
| Time to a complete answer | 108.1 s | – |
Each model is compared at its best-ranked reasoning-effort setting. Gemini 4 Carbon (high) scores 65.1 and takes 103 s to its first answer token. Muse Wonder scores 64.0 and takes 4 s to its first answer token; Muse Wonder (max) scores 62.1 in 33 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
92 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 | Gemini 4 Carbon (high) | Muse Wonder |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 65.1 | 64.0 |
| Coding score | BenchLeader | 71.7 | – |
| Knowledge score | BenchLeader | 75.8 | 76.4 |
| Long context score | BenchLeader | 64.5 | – |
| Reasoning score | BenchLeader | 78.9 | – |
| Agents & tools score | BenchLeader | – | 78.7 |
| Composite score | BenchLeader | – | 60.2 |
| Human preference score | BenchLeader | – | 69.6 |
| Maths score | BenchLeader | – | 70.3 |
| Blended price $/M | BenchLeader | $8.00 | $8.00 |
| Output speed | BenchLeader | 100 tok/s | 81 tok/s |
| Time to first answer | BenchLeader | 103.1 s | 3.6 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| Arena Hard Prompts | Arena (LMArena) | – | 1522† |
| LiveBench 2.0 Reasoning | LiveBench | – | 90.6% |
| LiveBench 1.0 Reasoning | LiveBench | – | 90.6%† |
| Humanity's Last Exam (AA) | Artificial Analysis | 60.4% | 50.1%† |
| HLE-Diamond (AA) | Artificial Analysis | 69.1% | 39.0%† |
| CritPt | Artificial Analysis | 36.0% | 28.3%† |
| CritPt v2 hard subset (AA) | Artificial Analysis | 63.0% | 48.0%† |
| MysteryMechanism | Vals AI | – | 27.5%† |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | – | 51.0%† |
| LMCA | Epoch AI Benchmarking Hub | – | 56.8% |
| DTBench | Epoch AI Benchmarking Hub | – | 95.2% |
| Coding | |||
| SciCode | SciCode | 64.2% | 60.3%† |
| Arena Coding | Arena (LMArena) | – | 1545† |
| Arena WebDev | Arena (LMArena) | – | 1652† |
| LiveBench 2.0 Coding | LiveBench | – | 86.2% |
| LiveBench 1.0 Coding | LiveBench | – | 80.7%† |
| IOI | Vals AI | – | 66.9%† |
| 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 | 89.1% | 83.1%† |
| CursorBench | Cursor | – | 50.4%† |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 65.2% | 52.0%† |
| Startrise Frontend | Startrise | – | 75.3% |
Show all 92 benchmarks
| Metric | Publisher | Gemini 4 Carbon (high) | Muse Wonder |
|---|---|---|---|
| Coding | |||
| FrontierCode 1.1 (AA) | Artificial Analysis | 48.5% | 46.5%† |
| FrontierCode 2.0 (AA) | Artificial Analysis | 40.6% | 27.9%† |
| Agents & tools | |||
| APEX-Agents | Mercor | – | 66.1%† |
| Arena Agent | Arena (LMArena) | – | 3.3† |
| LiveBench Long Agentic | LiveBench | – | 52.6% |
| LiveBench Browser | LiveBench | – | 60.0% |
| LiveBench Agentic Coding | LiveBench | – | 62.7%† |
| HiL-Bench | Scale AI SEAL | – | 79.7% |
| GDPval-AA v2.1 | Artificial Analysis | – | 63.5%† |
| CUA-bench | Vals AI | – | 7.2%† |
| CyberBench | Vals AI | – | 80.4%† |
| Finance Agent v2 | Vals AI | – | 56.6%† |
| Finance Agent v2.1 | Vals AI | 67.2% | 65.5%† |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 17.5% | 23.8%† |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 17.1%† |
| Legal Research Bench | Vals AI | 55.8% | 50.5%† |
| Tax Agent Bench | Vals AI | 48.9% | 36.3%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 63.1% | 49.5%† |
| Terminal-Bench Science | Vals AI | – | 22.9%† |
| Vending-Bench 2 | Andon Labs | 16484.2† | 9432.2 |
| AutomationBench | Artificial Analysis | 77.8% | 68.1%† |
| GDP.pdf (AA) | Artificial Analysis | 28.4% | 30.6%† |
| Harvey LAB-AA v1.1 | Harvey | – | 10.0%† |
| AA-Briefcase v1.1 | Artificial Analysis | 1609 | 1659† |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 64.8% | 25.7%† |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | – | 82.1%† |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | – | 70.7%† |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 99.4%† |
| ProofBench | Vals AI | – | 84.0% |
| LiveBench Mathematics | LiveBench | – | 96.5%† |
| Arena Maths | Arena (LMArena) | – | 1496† |
| Knowledge | |||
| LiveBench 2.0 Data Analysis | LiveBench | – | 81.6% |
| LiveBench Research | LiveBench | – | 88.0% |
| LiveBench 1.0 Data Analysis | LiveBench | – | 81.6%† |
| AA-Omniscience v1.1 | Artificial Analysis | 53.5 | 25.1† |
| PRBench Finance | Scale AI SEAL | – | 58.7% |
| PRBench Legal | Scale AI SEAL | – | 58.4% |
| BioMysteryBench | Vals AI | – | 67.0%† |
| Excel Modeling Benchmark | Vals AI | – | 67.1%† |
| Excel Modeling Benchmark v1.1 | Vals AI | 76.5% | 72.5%† |
| Public Benefits Bench v1.1 | Vals AI | – | 70.2%† |
| EBR-bench | Epoch AI Benchmarking Hub | – | 31.4%† |
| GDP.pdf | Epoch AI Benchmarking Hub | – | 30.8%† |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 58.3% | 38.0%† |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 88.6% | 79.1%† |
| Instruction following | |||
| LiveBench Language | LiveBench | – | 82.9%† |
| LiveBench Instruction Following | LiveBench | – | 73.0%† |
| Arena Instruction Following | Arena (LMArena) | – | 1498† |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1509† |
| EQ-Bench 4 | EQ-Bench | – | 1325 |
| Arena Creative Writing | Arena (LMArena) | – | 1492† |
| Arena Multi-turn | Arena (LMArena) | – | 1519† |
| Arena Longer Queries | Arena (LMArena) | – | 1502† |
| Multimodal | |||
| LiveBench Design | LiveBench | – | 75.1% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 58.6%† | – |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 79.3% | 80.7%† |
| MLCR | Artificial Analysis | – | 22.2%† |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 161.8 |
| LiveBench 2.0 | LiveBench | – | 74.2% |
| LiveBench 1.0 | LiveBench | – | 81.2%† |
| AA Intelligence Index v5.0 | Artificial Analysis | 59.4 | 43.3† |
| AA Intelligence Index v4.3.2 | Artificial Analysis | – | 52.4† |
| AA Progress Index v1 | Artificial Analysis | 351.6 | 335.8† |
| Vals Index | Vals AI | 70.4 | 64.9† |
| Vals Finance Index | Vals AI | 66.3% | 61.2%† |
| Vals Science Index | Vals AI | – | 39.1%† |
| Vals Router Index | Vals AI | – | 64.9% |
Data as of 2026-10-29. Best configuration of each model; every score links to its source on the model pages.
What changed
- Muse Wonder: 1 new result (Startrise Frontend).
- Muse Wonder: 3 score revisions.
- Muse Wonder: 1 new result (EQ-Bench 4).
- Muse Wonder: 1 new result (Vals Router Index).
- Muse Wonder: 1 new result (Epoch Capabilities Index).
- Muse Wonder: 10 new results (LMCA, DTBench, LiveBench 2.0…).
Related comparisons
- Gemini 4 Carbon vs Claude Opus 5.5
- Gemini 4 Carbon vs GPT-6 Astra
- Gemini 4 Carbon vs GPT-6.1 Sol
- Muse Wonder vs Claude Opus 5.5
- Muse Wonder vs GPT-6 Astra
- Muse Wonder vs GPT-6.1 Sol
- Muse Wonder vs Claude Fable 5.1
- Gemini 4 Argon vs Muse Wonder
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Gemini 4 Carbon vs Muse Wonder: questions
- Is Gemini 4 Carbon better than Muse Wonder?
- Gemini 4 Carbon (high) and Muse Wonder are neck and neck on quality (65.1 vs 64.0, within the ±7.9 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 29 Oct 2026, so check the category scores for your use.
- Gemini 4 Carbon vs Muse Wonder: which is cheaper, and will that change?
- They cost the same: $8.00 per million tokens blended. No price change has been announced for either.
- Which is faster, Gemini 4 Carbon or Muse Wonder?
- Gemini 4 Carbon streams faster (100 vs 81 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- Both take about 1M tokens: Gemini 4 Carbon 1,000,000, Muse Wonder 1,048,576.
- Which has lower latency?
- Muse Wonder answers first: 3.6 s against 103.1 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.
More questions (1)
- When was Gemini 4 Carbon released?
- Gemini 4 Carbon was released on 28 Oct 2026.