Muse Spark vs Step 5 Preview
Full profiles: Muse Spark (Meta) and Step 5 Preview (StepFun): every benchmark, price and rank.
Updated · Data from 8 sources
As of 16 Nov 2026, Muse Spark is ahead of Step 5 Preview on quality (64.2 vs 56.0 on the BenchLeader Index). Choose Muse Spark for agents & tools (65 vs 49), reasoning (67 vs 52), and coding (65 vs 50); choose Step 5 Preview for long context (69 vs 64) and maths (60 vs 59).
Key facts (as of 16 Nov 2026)
- Muse Spark scores 64.2 on the BenchLeader Index and Step 5 Preview scores 56.0 (BenchLeader, 16 Nov 2026).
- Step 5 Preview costs $1.43 per 1M tokens blended (list price, 16 Nov 2026).
- Of the 19 benchmarks both models have, Step 5 Preview leads 10 (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 50 category score; Arena Coding 1529 vs 1505 | High |
| Agents & tool use | Muse Spark | 65 vs 49 category score; GDPval-AA v2.1 1006 vs 1591 | High |
| Maths | Step 5 Preview | 60 vs 59 category score; Arena Maths 1470 vs 1469 | Medium |
| Reasoning | Muse Spark | 67 vs 52 category score; Arena Hard Prompts 1506 vs 1477 | High |
| Knowledge | Muse Spark | 63 vs 54 category score | High |
| Long documents | Step 5 Preview | 69 vs 64 category score; AA-LCR v1.1 88.3% vs 78.0% | Medium |
| Images and other media | Muse Spark | 64 vs 59 category score; Arena Vision 1306 vs 1263 | Medium |
| Chat quality (human preference) | Muse Spark | 68 vs 63 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 56.0.
- Muse Spark is stronger in agents & tools, coding, human preference, knowledge, multimodal, reasoning.
- Step 5 Preview is stronger in long context, maths.
At a glance
| Spec | Muse Spark | Step 5 Preview |
|---|---|---|
| Developer | Meta | StepFun |
| Released | 8 Apr 2026 | 18 Sep 2026 |
| Context window (tokens) | 262,144 | 1,000,000 |
| Compared at | Muse Spark | Step 5 Preview |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | – | $1.00 / $2.70 / $0.105 |
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 Spark | Step 5 Preview | Leader |
|---|---|---|---|
| Reasoning | 67.4 | 51.7 | Muse Spark |
| Coding | 65.1 | 49.9 | Muse Spark |
| Agents & tools | 65.0 | 49.1 | Muse Spark |
| Maths | 59.1 | 60.1 | Step 5 Preview |
| Knowledge | 63.1 | 53.7 | Muse Spark |
| Instruction following | 76.9 | – | not comparable |
| Human preference | 68.1 | 63.1 | Muse Spark |
| Multimodal | 64.2 | 59.3 | Muse Spark |
| Long context | 63.6 | 69.0 | Step 5 Preview |
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: Instruction following.
Where sources disagree
Across the 19 benchmarks both have, Muse Spark leads 9, Step 5 Preview leads 10. Every overall index that rates both puts Muse Spark ahead.
| Overall index | Muse Spark | Step 5 Preview | Leader |
|---|---|---|---|
| BenchLeader Index | 64.2 | 56.0 | Muse Spark |
| Arena Text | 1489 | 1449 | 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 | Muse Spark | Step 5 Preview |
|---|---|---|
| Blended price per 1M tokens | – | $1.43 |
| Input / output per 1M | – / – | $1.00 / $2.70 |
| Cached input per 1M | – | $0.105 |
| Price per task: chat reply | – | $0.0012 |
| Price per task: summarize a 30-page report | – | $0.014 |
| Price per task: code edit | – | $0.010 |
| Price per task: agentic coding session | – | $0.071 |
| Price per task: structured extraction | – | $0.0025 |
| Agentic coding session, input mostly cached | – | $0.030 |
| Cost per benchmark task (AA Index) | – | $2.35 |
| Output tokens per task (AA Index) | – | 113k |
| Of which reasoning | – | 80k |
| Time per task (AA Index) | – | 15 min |
| Output speed | – | 124 tok/s |
| Time to first answer | – | 23.5 s |
| Time to a complete answer, last measured 13 Nov 2026 | – | 54.4 s |
Each model is compared at its best-ranked reasoning-effort setting. Muse Spark scores 64.2. Step 5 Preview scores 56.0 and takes 24 s to its first answer token. 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
115 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 Spark | Step 5 Preview |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 64.2 | 56.0 |
| Agents & tools score | BenchLeader | 65.0 | 49.1 |
| Coding score | BenchLeader | 65.1 | 49.9 |
| Human preference score | BenchLeader | 68.1 | 63.1 |
| Instruction following score | BenchLeader | 76.9 | – |
| Knowledge score | BenchLeader | 63.1 | 53.7 |
| Long context score | BenchLeader | 63.6 | 69.0 |
| Maths score | BenchLeader | 59.1 | 60.1 |
| Multimodal score | BenchLeader | 64.2 | 59.3 |
| Reasoning score | BenchLeader | 67.4 | 51.7 |
| Blended price $/M | BenchLeader | – | $1.43 |
| Output speed | BenchLeader | – | 124 tok/s |
| Time to first answer | BenchLeader | – | 23.5 s |
| Context window | BenchLeader | 262,144 tokens | 1,000,000 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 | 1477 |
| LiveBench 2.0 Reasoning | LiveBench | – | 83.8%† |
| GPQA Diamond (AA) | Artificial Analysis | 88.4% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 40.7% | 46.5% |
| HLE-Diamond (AA) | Artificial Analysis | – | 24.3% |
| GPQA Diamond (Vals) | Vals AI | 89.7% | – |
| CritPt | Artificial Analysis | 11.3% | 20.9% |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 32.5% |
| MysteryMechanism | Vals AI | – | 18.5% |
| Arena Expert | Arena (LMArena) | 1491 | 1489 |
| Coding | |||
| SciCode | SciCode | 51.5% | 58.9% |
| Arena Coding | Arena (LMArena) | 1529 | 1505 |
| Arena WebDev | Arena (LMArena) | – | 1573† |
| LiveBench 2.0 Coding | LiveBench | – | 67.6%† |
| SciCode (AA) | Artificial Analysis | – | 58.9% |
| IOI | Vals AI | – | 38.1% |
| SWE-bench (Vals) | Vals AI | 74.4% | – |
| SWE-Bench Pro | Scale AI SEAL | 55.0% | – |
| Code Migration | Vals AI | – | 39.0% |
| Code Migration v1.1 | Vals AI | – | 47.0% |
| SWE-bench Multimodal | Vals AI | – | 40.8% |
| Vibe Code Bench v1.1 | Vals AI | 19.7% | 70.4% |
| SWE-Bench Pro (private) | Scale AI SEAL | 44.7% | – |
Show all 115 benchmarks
| Metric | Publisher | Muse Spark | Step 5 Preview |
|---|---|---|---|
| Coding | |||
| SWE Atlas v1.0: Codebase QnA | Scale AI SEAL | 24.2% | – |
| SWE Atlas v1.0: Test Writing | Scale AI SEAL | 31.1% | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | – | 33.3% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 62.2% | – |
| Design Arena Website | Design Arena | 1253 | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 38.4% |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 12.0% |
| Agents & tools | |||
| Arena Agent | Arena (LMArena) | – | -3.7 |
| LiveBench Long Agentic | LiveBench | – | 19.4%† |
| LiveBench Browser | LiveBench | – | 24.8%† |
| Terminal-Bench Hard | Artificial Analysis | 45.5% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 91.5% | – |
| MCP Atlas | Scale AI SEAL | 82.2% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | – | 54.3% |
| GDPval-AA v2.1 | Artificial Analysis | 1006 | 1591 |
| ITBench SRE (AA) | Artificial Analysis | – | 55.6% |
| Analyst Agent (AA) | Artificial Analysis | – | 35.0% |
| APEX-Agents (AA) | Artificial Analysis | – | 38.0% |
| Finance Agent v2 | Vals AI | – | 50.7% |
| Finance Agent v2.1 | Vals AI | – | 58.9% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 12.9% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 9.2% |
| Legal Research Bench | Vals AI | – | 40.4% |
| SREBench | Vals AI | – | 0.4% |
| Tax Agent Bench | Vals AI | – | 27.3% |
| Terminal-Bench 2.0 (Vals) | Vals AI | 59.5% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | – | 31.8% |
| Terminal-Bench Science | Vals AI | – | 2.9% |
| Arena Agent Task Completion | Arena (LMArena) | – | -1.5 |
| AutomationBench | Artificial Analysis | – | 51.0% |
| GDP.pdf (AA) | Artificial Analysis | – | 14.8% |
| EnterpriseOps-Gym | Artificial Analysis | – | 47.2% |
| AA-Briefcase v1.1 | Artificial Analysis | 646 | 1424 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 2.4% |
| CWE-Bench (AA) | Artificial Analysis | – | 52.5% |
| DeepSec-Bench (AA) | Artificial Analysis | – | 15.8% |
| CyberGym end-to-end (AA) | Artificial Analysis | – | 74.8% |
| MKA (AA) | Artificial Analysis | – | 27.6% |
| Maths | |||
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 88.9% | – |
| ProofBench | Vals AI | 17.0% | 42.0% |
| AIME (Vals) | Vals AI | 96.9% | – |
| Arena Maths | Arena (LMArena) | 1469 | 1470 |
| Knowledge | |||
| LiveBench 2.0 Data Analysis | LiveBench | – | 63.5%† |
| LiveBench Research | LiveBench | – | 75.5%† |
| AA-Omniscience v1.1 | Artificial Analysis | – | 16.4 |
| 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% | – |
| BioMysteryBench | Vals AI | – | 67.4% |
| CaseLaw v2 | Vals AI | 63.1% | – |
| Excel Modeling Benchmark | Vals AI | – | 57.1% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 69.5% |
| MedCode | Vals AI | 51.3% | – |
| MedScribe | Vals AI | 85.9% | – |
| Public Benefits Bench v1.1 | Vals AI | – | 58.1% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | – | 36.4% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | – | 68.7% |
| 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 | 1450 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1489 | 1449 |
| Arena Creative Writing | Arena (LMArena) | 1464 | 1413 |
| Arena Multi-turn | Arena (LMArena) | 1493 | 1445 |
| Arena Longer Queries | Arena (LMArena) | 1476 | 1461 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1306 | 1263 |
| LiveBench Design | LiveBench | – | 70.8%† |
| MMMU-Pro | Artificial Analysis | 80.5% | 76.4% |
| MMMU-Pro (Vals) | Vals AI | 87.4% | – |
| Code2Video-AA | Artificial Analysis | – | 821 |
| AA-3D | Artificial Analysis | – | 859 |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 78.0% | 88.3% |
| Arena Document | Arena (LMArena) | 1467 | – |
| MLCR | Artificial Analysis | – | 18.3% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 153.1 | – |
| LiveBench 2.0 | LiveBench | – | 54.9%† |
| AA Intelligence Index v5.0.1 | Artificial Analysis | – | 28.5 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 31.3 | 43.7 |
| AA Progress Index v1 | Artificial Analysis | 289.1 | 318.2 |
| Vals Index | Vals AI | – | 54.8 |
| Vals Finance Index | Vals AI | – | 55.4% |
| Vals Science Index | Vals AI | – | 29.6% |
| Vals Legal Index | Vals AI | – | 70.2% |
| AA Cyber Index | Artificial Analysis | – | 47.7 |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 20.2% | – |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 33.3 |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
What changed
- Step 5 Preview: BenchLeader Index 54.2 → 56.0.
- Step 5 Preview: BenchLeader Index 54.7 → 54.2.
- Muse Spark: 2 new results (GDPval-AA v2.1, AA-Briefcase v1.1).
- Step 5 Preview: 1 score revision.
- Step 5 Preview: 7 score revisions.
- Muse Spark: BenchLeader Index 64.6 → 64.1.
- Muse Spark: 1 score revision.
- Step 5 Preview: 2 new results (Vals Index, Harvey's Legal Agent Benchmark v1.1).
- Step 5 Preview: 1 score revision.
- Step 5 Preview: BenchLeader Index 58.0 → 54.7.
Related comparisons
- Muse Spark vs Claude Fable 5.2
- Muse Spark vs Gemini 4 Carbon
- Muse Spark vs Claude Opus 5.5
- Step 5 Preview vs Claude Opus 5.5
- Step 5 Preview vs GPT-6 Astra
- Step 5 Preview vs GPT-6.1 Sol
- Step 5 Preview vs Gemini 4 Argon
- Muse Spark vs Claude Fable 5.1
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Muse Spark vs Step 5 Preview: questions
- Is Muse Spark better than Step 5 Preview?
- Muse Spark leads on quality: 64.2 vs 56.0. 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 Muse Spark or Step 5 Preview better for coding?
- Muse Spark scores higher in coding (65 vs 50 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 51.5% for Muse Spark against 58.9% for Step 5 Preview.
- Is Muse Spark or Step 5 Preview better for agents and tool use?
- Muse Spark scores higher in agentic tasks (65 vs 49 on the category index, where 50 is average).
- Which is better for long documents?
- Step 5 Preview scores higher on long-context benchmarks (69 vs 64). Step 5 Preview accepts more context: 1,000,000 against 262,144 tokens for Muse Spark.
- Do independent benchmarks agree on Muse Spark vs Step 5 Preview?
- Across the 19 benchmarks both have, Muse Spark leads 9, Step 5 Preview leads 10. Every overall index that rates both puts Muse Spark ahead.
More questions (1)
- When was Step 5 Preview released?
- Step 5 Preview was released on 18 Sep 2026.