Gemini 4 Carbon vs Muse Spark
Full profiles: Gemini 4 Carbon (Google) and Muse Spark (Meta): every benchmark, price and rank.
Updated · Data from 5 sources
As of 29 Oct 2026, Gemini 4 Carbon and Muse Spark are neck and neck on quality (65.1 vs 64.3 on the BenchLeader Index, inside the ±7.0 margin). Gemini 4 Carbon leads 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 Spark scores 64.3 (BenchLeader, 29 Oct 2026).
- Gemini 4 Carbon costs $8.00 per 1M tokens blended (list price, 29 Oct 2026).
- Of the 6 benchmarks both models have, Gemini 4 Carbon leads 6 (BenchLeader, 29 Oct 2026).
Gemini 4 Carbon benchmarks 
Muse Spark 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 |
|---|---|---|---|
| Coding | Gemini 4 Carbon | 72 vs 65 category score; SciCode 64.2% vs 51.5% | Medium |
| Reasoning | Gemini 4 Carbon | 79 vs 68 category score; Humanity's Last Exam (AA) 60.4% vs 40.7% | High |
| Knowledge | Gemini 4 Carbon | 76 vs 63 category score | High |
| Long documents | Either | 65 vs 64 category score; AA-LCR v1.1 79.3% vs 78.0% | Either |
| Chat quality (human preference) | – | Only Muse Spark has an Arena rating (68 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
- Gemini 4 Carbon (high) and Muse Spark are neck and neck on quality (65.1 vs 64.3, within the ±7.0 margin).
- Gemini 4 Carbon (high) is stronger in coding, knowledge, long context, reasoning.
- Muse Spark is stronger in agents & tools, human preference, instruction following, maths, multimodal.
At a glance
| Spec | Gemini 4 Carbon | Muse Spark |
|---|---|---|
| Developer | Meta | |
| Released | 28 Oct 2026 | 8 Apr 2026 |
| Context window (tokens) | 1,000,000 | 262,144 |
| Compared at | Gemini 4 Carbon (high) | Muse Spark |
| Open weights | No | No |
| Price in / out / cached input per 1M | $4.00 / $20.00 / – | – |
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 Spark | Leader |
|---|---|---|---|
| Reasoning | 78.9 | 67.6 | Gemini 4 Carbon |
| Coding | 71.7 | 64.9 | Gemini 4 Carbon |
| Agents & tools | – | 66.1 | not comparable |
| Maths | – | 59.0 | not comparable |
| Knowledge | 75.8 | 62.9 | Gemini 4 Carbon |
| Instruction following | – | 76.6 | not comparable |
| Human preference | – | 68.2 | not comparable |
| Multimodal | – | 64.3 | not comparable |
| Long context | 64.5 | 63.8 | level |
Where sources disagree
Across the 6 benchmarks both have, Gemini 4 Carbon leads 6, Muse Spark leads 0.
| Overall index | Gemini 4 Carbon (high) | Muse Spark | Leader |
|---|---|---|---|
| BenchLeader Index | 65.1 | 64.3 | neck and neck |
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 Spark |
|---|---|---|
| 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: summarize 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 | $7.76 | – |
| Output speed | 100 tok/s | – |
| Time to first answer | 103.1 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 Spark scores 64.3. 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
76 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 Spark |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 65.1 | 64.3 |
| Coding score | BenchLeader | 71.7 | 64.9 |
| Knowledge score | BenchLeader | 75.8 | 62.9 |
| Long context score | BenchLeader | 64.5 | 63.8 |
| Reasoning score | BenchLeader | 78.9 | 67.6 |
| Agents & tools score | BenchLeader | – | 66.1 |
| Human preference score | BenchLeader | – | 68.2 |
| Instruction following score | BenchLeader | – | 76.6 |
| Maths score | BenchLeader | – | 59.0 |
| Multimodal score | BenchLeader | – | 64.3 |
| Blended price $/M | BenchLeader | $8.00 | – |
| Output speed | BenchLeader | 100 tok/s | – |
| Time to first answer | BenchLeader | 103.1 s | – |
| Context window | BenchLeader | 1,000,000 tokens | 262,144 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 89.8% |
| Humanity's Last Exam | Scale AI / CAIS | – | 40.6% |
| Arena Hard Prompts | Arena (LMArena) | – | 1506 |
| GPQA Diamond (AA) | Artificial Analysis | – | 88.4% |
| Humanity's Last Exam (AA) | Artificial Analysis | 60.4% | 40.7% |
| HLE-Diamond (AA) | Artificial Analysis | 69.1% | – |
| GPQA Diamond (Vals) | Vals AI | – | 89.7% |
| CritPt | Artificial Analysis | 36.0% | 11.3% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 63.0% | – |
| Coding | |||
| SciCode | SciCode | 64.2% | 51.5% |
| Arena Coding | Arena (LMArena) | – | 1529 |
| SciCode (AA) | Artificial Analysis | 64.2% | – |
| SWE-bench (Vals) | Vals AI | – | 74.4% |
| SWE-Bench Pro | Scale AI SEAL | – | 55.0% |
| Vibe Code Bench v1.1 | Vals AI | 89.1% | 19.7% |
| SWE-Bench Pro (private) | Scale AI SEAL | – | 55.0% |
| SWE Atlas: Codebase QnA | Scale AI SEAL | – | 24.2% |
| SWE Atlas: Test Writing | Scale AI SEAL | – | 31.1% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 65.2% | – |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 62.2% |
| FrontierCode 1.1 (AA) | Artificial Analysis | 48.5% | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 40.6% | – |
| Agents & tools | |||
| Terminal-Bench Hard | Artificial Analysis | – | 45.5% |
| τ²-Bench Telecom (AA) | Artificial Analysis | – | 91.5% |
| MCP Atlas | Scale AI SEAL | – | 82.2% |
Show all 76 benchmarks
| Metric | Publisher | Gemini 4 Carbon (high) | Muse Spark |
|---|---|---|---|
| Agents & tools | |||
| Finance Agent v2.1 | Vals AI | 67.2% | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 17.5% | – |
| Legal Research Bench | Vals AI | 55.8% | – |
| Tax Agent Bench | Vals AI | 48.9% | – |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 59.5% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 63.1% | – |
| Vending-Bench 2 | Andon Labs | 16484.2† | – |
| AutomationBench | Artificial Analysis | 77.8% | – |
| GDP.pdf (AA) | Artificial Analysis | 28.4% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1609 | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 64.8% | – |
| Maths | |||
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 88.9% |
| ProofBench | Vals AI | – | 17.0% |
| AIME (Vals) | Vals AI | – | 96.9% |
| Arena Maths | Arena (LMArena) | – | 1467 |
| Knowledge | |||
| AA-Omniscience v1.1 | Artificial Analysis | 53.5 | – |
| MMLU-Pro | Vals AI | – | 87.3% |
| LegalBench | Vals AI | – | 84.2% |
| CorpFin | Vals AI | – | 65.1% |
| TaxEval | Vals AI | – | 77.7% |
| PRBench Finance | Scale AI SEAL | – | 52.4% |
| PRBench Legal | Scale AI SEAL | – | 52.3% |
| MultiNRC | Scale AI SEAL | – | 59.0% |
| CaseLaw v2 | Vals AI | – | 63.1% |
| Excel Modeling Benchmark v1.1 | Vals AI | 76.5% | – |
| MedCode | Vals AI | – | 51.3% |
| MedScribe | Vals AI | – | 85.9% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 58.3% | – |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 88.6% | – |
| Instruction following | |||
| IFBench | Artificial Analysis | – | 75.9% |
| MultiChallenge | Scale AI SEAL | – | 75.5% |
| TutorBench | Scale AI SEAL | – | 68.5% |
| Arena Instruction Following | Arena (LMArena) | – | 1465 |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1489 |
| Arena Creative Writing | Arena (LMArena) | – | 1464 |
| Arena Multi-turn | Arena (LMArena) | – | 1494 |
| Arena Longer Queries | Arena (LMArena) | – | 1476 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | – | 1306 |
| MMMU-Pro | Artificial Analysis | – | 80.5% |
| MMMU-Pro (official) | MMMU | – | 80.4%† |
| MMMU-Pro (Vals) | Vals AI | – | 87.4% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 58.6%† | – |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 79.3% | 78.0% |
| Arena Document | Arena (LMArena) | – | 1467 |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 152.0 |
| AA Intelligence Index v5.0 | Artificial Analysis | 59.4 | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | – | 31.3 |
| AA Progress Index v1 | Artificial Analysis | 351.6 | 289.1 |
| Vals Index | Vals AI | 70.4 | – |
| Vals Finance Index | Vals AI | 66.3% | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | – | 20.2% |
Data as of 2026-10-29. Best configuration of each model; every score links to its source on the model pages.
What changed
- Muse Spark: 1 new result (AA Progress Index v1).
- Muse Spark: 1 new result (PRBench Legal).
- Muse Spark: 1 new result (SWE Atlas: Codebase QnA).
- Muse Spark: 2 new results (PRBench Legal, MultiNRC).
- Muse Spark: BenchLeader Index 65.1 → 64.6.
- Muse Spark: 1 new result (PRBench Finance).
- Muse Spark: 1 new result (PRBench Finance).
- Muse Spark: BenchLeader Index 64.6 → 65.1.
- Muse Spark: 1 new result (TutorBench); 1 score revision.
- Muse Spark: BenchLeader Index 65.3 → 64.6.
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 Spark vs Claude Opus 5.5
- Muse Spark vs GPT-6 Astra
- Muse Spark vs GPT-6.1 Sol
- Muse Spark vs Claude Fable 5.1
- Gemini 4 Argon vs Muse Spark
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Gemini 4 Carbon vs Muse Spark: questions
- Is Gemini 4 Carbon better than Muse Spark?
- Gemini 4 Carbon (high) and Muse Spark are neck and neck on quality (65.1 vs 64.3, within the ±7.0 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.
- Is Gemini 4 Carbon or Muse Spark better for coding?
- Gemini 4 Carbon scores higher in coding (72 vs 65 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 64.2% for Gemini 4 Carbon against 51.5% for Muse Spark.
- Which is better for long documents?
- They are level on long-context benchmarks (65 vs 64). Gemini 4 Carbon accepts more context: 1,000,000 against 262,144 tokens for Muse Spark.
- Do independent benchmarks agree on Gemini 4 Carbon vs Muse Spark?
- Across the 6 benchmarks both have, Gemini 4 Carbon leads 6, Muse Spark leads 0.
- When was Gemini 4 Carbon released?
- Gemini 4 Carbon was released on 28 Oct 2026.