GPT-5.1 vs Muse Spark
Full profiles: GPT-5.1 (OpenAI) and Muse Spark (Meta): every benchmark, price and rank.
Updated · Data from 10 sources
As of 16 Nov 2026, Muse Spark is ahead of GPT-5.1 on quality (64.2 vs 57.5 on the BenchLeader Index). Muse Spark leads in every category we measure both in.
Key facts (as of 16 Nov 2026)
- GPT-5.1 scores 57.5 on the BenchLeader Index and Muse Spark scores 64.2 (BenchLeader, 16 Nov 2026).
- GPT-5.1 costs $3.44 per 1M tokens blended (list price, 16 Nov 2026).
- Of the 37 benchmarks both models have, Muse Spark leads 30 (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 |
|---|---|---|---|
| Coding | Muse Spark | 65 vs 56 category score; Arena Coding 1529 vs 1492 | High |
| Agents & tool use | Muse Spark | 65 vs 41 category score; MCP Atlas 82.2% vs 50.1% | High |
| Maths | Either | 59 vs 59 category score; Arena Maths 1459 vs 1469 | Either |
| Reasoning | Muse Spark | 67 vs 53 category score; GPQA Diamond 89.8% vs 87.6% | High |
| Knowledge | Muse Spark | 63 vs 59 category score; MMLU-Pro 87.3% vs 86.4% | Medium |
| Long documents | Either | 65 vs 64 category score; AA-LCR v1.1 80.0% vs 78.0% | Either |
| Images and other media | Muse Spark | 64 vs 58 category score; Arena Vision 1306 vs 1249 | Medium |
| Following instructions | Muse Spark | 77 vs 71 category score; IFBench 75.9% vs 72.9% | Medium |
| Chat quality (human preference) | Muse Spark | 68 vs 64 on Arena votes | Medium |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Muse Spark leads on quality: 64.2 vs 57.5.
- GPT-5.1 (high) is stronger in long context.
- Muse Spark is stronger in agents & tools, coding, human preference, instruction following, knowledge, maths, multimodal, reasoning.
At a glance
| Spec | GPT-5.1 | Muse Spark |
|---|---|---|
| Developer | OpenAI | Meta |
| Released | 12 Nov 2025 | 8 Apr 2026 |
| Context window (tokens) | 400,000 | 262,144 |
| Compared at | GPT-5.1 (high) | Muse Spark |
| Open weights | No | No |
| Price in / out / cached input per 1M | $1.25 / $10.00 / $0.125 | – |
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-5.1 (high) | Muse Spark | Leader |
|---|---|---|---|
| Reasoning | 53.0 | 67.4 | Muse Spark |
| Coding | 55.7 | 65.1 | Muse Spark |
| Agents & tools | 41.2 | 65.0 | Muse Spark |
| Maths | 58.5 | 59.1 | level |
| Knowledge | 58.6 | 63.1 | Muse Spark |
| Instruction following | 71.4 | 76.9 | Muse Spark |
| Human preference | 64.0 | 68.1 | Muse Spark |
| Multimodal | 57.9 | 64.2 | Muse Spark |
| Long context | 64.6 | 63.6 | level |
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.
Where sources disagree
Across the 37 benchmarks both have, GPT-5.1 leads 6, Muse Spark leads 30 and 1 is a tie. Every overall index that rates both puts Muse Spark ahead.
| Overall index | GPT-5.1 (high) | Muse Spark | Leader |
|---|---|---|---|
| BenchLeader Index | 57.5 | 64.2 | Muse Spark |
| Arena Text | 1455 | 1489 | Muse Spark |
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-5.1 (high) | Muse Spark |
|---|---|---|
| Blended price per 1M tokens | $3.44 | – |
| Input / output per 1M | $1.25 / $10.00 | – / – |
| Cached input per 1M | $0.125 | – |
| Price per task: chat reply | $0.0035 | – |
| Price per task: summarize a 30-page report | $0.021 | – |
| Price per task: code edit | $0.022 | – |
| Price per task: agentic coding session | $0.115 | – |
| Price per task: structured extraction | $0.0045 | – |
| Agentic coding session, input mostly cached | $0.064 | – |
| Output speed | 88 tok/s | – |
| Time to first answer | 36.8 s | – |
| Time to a complete answer, last measured 13 Nov 2026 | 80.1 s | – |
Each model is compared at its best-ranked reasoning-effort setting. GPT-5.1 (high) scores 57.5 and takes 37 s to its first answer token; GPT-5.1 scores 56.1 in 1 s, GPT-5.1 (thinking) scores 54.6 in 1 s. Muse Spark scores 64.2. 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
98 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-5.1 (high) | Muse Spark |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 57.5 | 64.2 |
| Agents & tools score | BenchLeader | 41.2 | 65.0 |
| Coding score | BenchLeader | 55.7 | 65.1 |
| Human preference score | BenchLeader | 64.0 | 68.1 |
| Instruction following score | BenchLeader | 71.4 | 76.9 |
| Knowledge score | BenchLeader | 58.6 | 63.1 |
| Long context score | BenchLeader | 64.6 | 63.6 |
| Maths score | BenchLeader | 58.5 | 59.1 |
| Multimodal score | BenchLeader | 57.9 | 64.2 |
| Reasoning score | BenchLeader | 53.0 | 67.4 |
| Blended price $/M | BenchLeader | $3.44 | – |
| Output speed | BenchLeader | 88 tok/s | – |
| Time to first answer | BenchLeader | 36.8 s | – |
| Context window | BenchLeader | 400,000 tokens | 262,144 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 87.6% | 89.8% |
| Humanity's Last Exam | Scale AI / CAIS | 23.7%† | 40.6% |
| SimpleBench | SimpleBench | 53.2% | – |
| Arena Hard Prompts | Arena (LMArena) | 1476 | 1506 |
| GPQA Diamond (AA) | Artificial Analysis | 87.3% | 88.4% |
| Humanity's Last Exam (AA) | Artificial Analysis | 28.5% | 40.7% |
| GPQA Diamond (Vals) | Vals AI | 86.6% | 89.7% |
| GPQA Diamond (HELM) | HELM Capabilities | 44.2%† | – |
| ARC-AGI-1 | ARC Prize | 72.8% | – |
| ARC-AGI-2 | ARC Prize | 17.6% | – |
| CritPt | Artificial Analysis | 4.9% | 11.3% |
| Arena Expert | Arena (LMArena) | 1481 | 1491 |
| Chess Puzzles | Epoch AI Benchmarking Hub | 42.0% | – |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 16.0%† | – |
| CL-bench | Epoch AI Benchmarking Hub | 23.7% | – |
| CL-bench Life | Epoch AI Benchmarking Hub | 17.3% | – |
| LMCA | Epoch AI Benchmarking Hub | 43.9% | – |
| DTBench | Epoch AI Benchmarking Hub | 90.1% | – |
| ForecastBench | ForecastBench | 62.0%† | – |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | 68.0% | – |
| SciCode | SciCode | 43.3%† | 51.5% |
| WeirdML | WeirdML | 60.8% | – |
| GSO-Bench | GSO-Bench | 13.7% | – |
| Arena Coding | Arena (LMArena) | 1492 | 1529 |
| Arena WebDev | Arena (LMArena) | 1342† | – |
Show all 98 benchmarks
| Metric | Publisher | GPT-5.1 (high) | Muse Spark |
|---|---|---|---|
| Coding | |||
| LiveCodeBench | Vals AI | 86.5% | – |
| SWE-bench (Vals) | Vals AI | 69.8% | 74.4% |
| SWE-Bench Pro | Scale AI SEAL | – | 55.0% |
| SWE-bench Verified (bash only) | SWE-bench | 66.0%† | – |
| SWE-bench Verified (any scaffold) | SWE-bench | 66.0%† | – |
| Vibe Code Bench v1.1 | Vals AI | 24.6% | 19.7% |
| SWE-Bench Pro (private) | Scale AI SEAL | – | 44.7% |
| SWE Atlas v1.0: Codebase QnA | Scale AI SEAL | – | 24.2% |
| SWE Atlas v1.0: Test Writing | Scale AI SEAL | – | 31.1% |
| ALE-Bench | Epoch AI Benchmarking Hub | 1192.2 | – |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 52.4% | 62.2% |
| Design Arena Website | Design Arena | 1179 | 1253 |
| LiveCodeBench (AA) | Artificial Analysis | 86.8% | – |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 47.6%† | – |
| Terminal-Bench Hard | Artificial Analysis | 45.5% | 45.5% |
| τ²-Bench Telecom (AA) | Artificial Analysis | 81.9% | 91.5% |
| MCP Atlas | Scale AI SEAL | 50.1% | 82.2% |
| GDPval-AA v2.1 | Artificial Analysis | 834 | 1006 |
| τ³-Banking (AA) | Artificial Analysis | 15.9% | – |
| Poker Agent | Vals AI | 1038.6%† | – |
| Terminal-Bench 2.0 (Vals) | Vals AI | 44.9% | 59.5% |
| Arena Search | Arena (LMArena) | 1184† | – |
| DeepResearch Bench | DeepResearch Bench | 42.8%† | – |
| Vending-Bench 2 | Andon Labs | 1473.4† | – |
| AA-Briefcase v1.1 | Artificial Analysis | – | 646 |
| Design Arena Agentic | Design Arena | 1009† | – |
| Maths | |||
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 88.6% | 88.9% |
| ProofBench | Vals AI | – | 17.0% |
| AIME (Vals) | Vals AI | 93.3% | 96.9% |
| MGSM | Vals AI | 93.0% | – |
| MathArena Apex | MathArena | 1.0% | – |
| Omni-MATH (HELM) | HELM Capabilities | 46.4%† | – |
| Arena Maths | Arena (LMArena) | 1459 | 1469 |
| AIME 2025 (AA) | Artificial Analysis | 94.0% | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 48.0% | – |
| MMLU-Pro | Vals AI | 86.4% | 87.3% |
| LegalBench | Vals AI | 85.7% | 84.2% |
| CorpFin | Vals AI | 63.8% | 65.1% |
| TaxEval | Vals AI | 74.9% | 77.7% |
| MedQA | Vals AI | 96.4% | – |
| PRBench Finance | Scale AI SEAL | 48.0%† | 52.4% |
| PRBench Legal | Scale AI SEAL | 49.3%† | 52.3% |
| MultiNRC | Scale AI SEAL | 49.0%† | 59.0% |
| MMLU-Pro (HELM) | HELM Capabilities | 57.9%† | – |
| CaseLaw v2 | Vals AI | 73.4% | 63.1% |
| MedCode | Vals AI | 52.7% | 51.3% |
| MedScribe | Vals AI | 88.1% | 85.9% |
| Global-MMLU-Lite (AA) | Artificial Analysis | 90.6% | – |
| Instruction following | |||
| IFBench | Artificial Analysis | 72.9% | 75.9% |
| MultiChallenge | Scale AI SEAL | 63.4%† | 75.5% |
| TutorBench | Scale AI SEAL | 54.1%† | 68.5% |
| IFEval (HELM) | HELM Capabilities | 93.5%† | – |
| WildBench (HELM) | HELM Capabilities | 86.3%† | – |
| Arena Instruction Following | Arena (LMArena) | 1450 | 1465 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1455 | 1489 |
| Arena Creative Writing | Arena (LMArena) | 1428 | 1464 |
| Arena Multi-turn | Arena (LMArena) | 1464 | 1493 |
| Arena Longer Queries | Arena (LMArena) | 1460 | 1476 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1249 | 1306 |
| MMMU-Pro | Artificial Analysis | 75.5% | 80.5% |
| VISTA | Scale AI SEAL | 43.8%† | – |
| MMMU-Pro (Vals) | Vals AI | 83.2% | 87.4% |
| MortgageTax | Vals AI | 61.4% | – |
| VPCT | Epoch AI Benchmarking Hub | 58.7% | – |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 80.0% | 78.0% |
| Arena Document | Arena (LMArena) | 1416† | 1467 |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 150.9† | 153.1 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 24.7 | 31.3 |
| AA Progress Index v1 | Artificial Analysis | 275.7 | 289.1 |
| HELM Capabilities mean | HELM Capabilities | 65.6%† | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 25.7%† | 20.2% |
| MASK | Scale AI SEAL | 86.3%† | – |
| 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.
What changed
- GPT-5.1 (high): 2 new results (GDPval-AA v2.1, AIME 2025 (AA)).
- Muse Spark: 2 new results (GDPval-AA v2.1, AA-Briefcase v1.1).
- Muse Spark: BenchLeader Index 64.6 → 64.1.
- Muse Spark: 1 score revision.
- GPT-5.1 (high): 1 new result (AA Progress Index v1).
- Muse Spark: 1 new result (AA Progress Index v1).
- GPT-5.1 (high): 1 new result (Global-MMLU-Lite (AA)).
- GPT-5.1 (high): BenchLeader Index 58.4 → 57.9.
Related comparisons
- Muse Spark vs Claude Fable 5.2
- Muse Spark vs Gemini 4 Carbon
- Muse Spark vs Claude Opus 5.5
- Muse Spark vs GPT-6 Astra
- Muse Spark vs Claude Fable 5.1
- Muse Spark vs GPT-6.1 Sol
- Muse Spark vs Claude Fable 5
- Muse Spark vs GPT-5.6 Sol
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
GPT-5.1 vs Muse Spark: questions
- Is Muse Spark better than GPT-5.1?
- Muse Spark leads on quality: 64.2 vs 57.5. The BenchLeader Index combines every independent quality benchmark; Muse Spark is ahead overall as of 16 Nov 2026, but check the category scores for your use.
- Is GPT-5.1 or Muse Spark better for coding?
- Muse Spark scores higher in coding (65 vs 56 on the category index, where 50 is average). On Arena Coding, the top shared coding benchmark, it is 1492 for GPT-5.1 against 1529 for Muse Spark.
- Is GPT-5.1 or Muse Spark better for agents and tool use?
- Muse Spark scores higher in agentic tasks (65 vs 41 on the category index, where 50 is average). On MCP Atlas, the top shared agents & tools benchmark, it is 50.1% for GPT-5.1 against 82.2% for Muse Spark.
- Which is better for long documents?
- They are level on long-context benchmarks (65 vs 64). GPT-5.1 accepts more context: 400,000 against 262,144 tokens for Muse Spark.
- Do independent benchmarks agree on GPT-5.1 vs Muse Spark?
- Across the 37 benchmarks both have, GPT-5.1 leads 6, Muse Spark leads 30 and 1 is a tie. Every overall index that rates both puts Muse Spark ahead.
More questions (1)
- When was Muse Spark released?
- Muse Spark was released on 8 Apr 2026.