Muse Wonder vs Step 5 Preview
Muse Wonder leads Step 5 Preview on coding, 71.9 to 61.1 on the Vibe Coding Index, and Step 5 Preview costs 7.4x less per output token ($2.70 vs $20.00 per million). Updated October 21, 2026; 44 benchmarks have a result for both.
Muse Wonder or Step 5 Preview: which to pick
Pick Muse Wonder for coding: it scores 71.9 on the Vibe Coding Index against 61.1 for Step 5 Preview, though on 3 benchmarks against 4.
Pick Step 5 Preview for output tokens 7.4x cheaper ($2.70 vs $20.00 per million) and work measured by BioMysteryBench (Vals AI) or Long context reasoning (AA-LCR), where it leads.
- Shared results
- 44
- Muse Wonder only
- 16
- Step 5 Preview only
- 17
- Like-for-like categories
- 6 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
Muse Wonder
Muse Wonder scores 71.9 against 61.1.
3 and 4 benchmarks behind the scores
Intelligence
AA Intelligence Index
No call
Only Step 5 Preview has a AA Intelligence Index result, 43.7.
Directional only
Web design
Design Arena: websites
No call
Only Muse Wonder has a Design Arena: websites result, 1362.
Directional only
General
Vals Index
No call
Only Muse Wonder has a Vals Index result, 64.9%.
Directional only
Long documents
Listed context window
No call
Both take 1.05M tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Step 5 Preview
Muse Wonder $0.014, Step 5 Preview $0.0024 per turn.
List prices
At a glance
- Coding index
- Muse Wonder 71.9 from 3 benchmarks, Step 5 Preview 61.1 from 4 benchmarks.
- Widest category
- Math, 18.6 points apart on average over 3 shared benchmarks.
- Widest single gap
- ProofBench: Muse Wonder 84.0% against Step 5 Preview 42.0%, 42.0 points apart on a 0 to 100 scale.
- Output price
- Muse Wonder $20.00 against Step 5 Preview $2.70 per million output tokens, 7.4x apart.
- Context window
- Both take 1.05M tokens.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Muse Wonder | Bars | Step 5 Preview | Ties |
|---|---|---|---|---|
| Coding | 11 | 0 | ||
| Agentic work | 6 | 1 | ||
| Reasoning | 5 | 0 | ||
| Math | 3 | 0 | ||
| Knowledge | 4 | 0 | ||
| General | 13 | 1 | ||
| All shared boards | 42 | 2 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index110
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| LiveBench coding | Not measured | 85.6% | Only Step 5 Preview measured |
| LMArena WebDev | 1652 | 1567 | Muse Wonder |
| Terminal-Bench 4 (Vals AI) | 49.5% | 31.8% | Muse Wonder |
| Vibe Code Bench (Vals AI) | 83.1% | 70.4% | Muse Wonder |
| Code migration (Vals AI) | 62.5% | 39.0% | Muse Wonder |
| IOI (Vals AI) | 66.9% | 38.1% | Muse Wonder |
| SWE-bench Multimodal (Vals AI) | 45.6% | 40.8% | Muse Wonder |
| Vibe Code Bench 1-100 (Vals AI) | 26.6% | Not measured | Only Muse Wonder measured |
| CyberBench (Vals AI) | 80.4% | Not measured | Only Muse Wonder measured |
| SciCode | 60.3% | 58.9% | Muse Wonder |
| LMArena Text: coding | 1522 | 1485 | Muse Wonder |
| LMArena Text: software and IT | 1515 | 1473 | Muse Wonder |
| LMArena WebDev: React | 1641 | 1583 | Muse Wonder |
| LMArena WebDev: HTML | 1694 | 1556 | Muse Wonder |
| LMArena WebDev: from an image | Not measured | 1550 | Only Step 5 Preview measured |
| Neither model has a result on the other 24 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| AA Intelligence Index | Not measured | 43.7 | Only Step 5 Preview measured |
| Epoch Capabilities Index | 161.8 | Not measured | Only Muse Wonder measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index61
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| Design Arena: spreadsheets | 1227 | Not measured | Only Muse Wonder measured |
| LiveBench research | Not measured | 72.8% | Only Step 5 Preview measured |
| LiveBench browser | Not measured | 24.0% | Only Step 5 Preview measured |
| LiveBench long agentic | Not measured | 27.5% | Only Step 5 Preview measured |
| SREBench (Vals AI) | Not measured | 0.4% | Only Step 5 Preview measured |
| Finance Agent (Vals AI) | 56.6% | 50.7% | Muse Wonder |
| Tax Agent Bench (Vals AI) | 36.3% | 27.3% | Muse Wonder |
| Legal Research Bench (Vals AI) | 50.5% | 40.4% | Muse Wonder |
| Harvey's Legal Agent Benchmark (Vals AI) | 23.8% | Not measured | Only Muse Wonder measured |
| Excel Modeling Benchmark (Vals AI) | 67.1% | 57.1% | Muse Wonder |
| BioMysteryBench (Vals AI) | 67.0% | 67.4% | Step 5 Preview |
| Terminal-Bench-Science (Vals AI) | 22.9% | 2.9% | Muse Wonder |
| Vals Science Index | 39.1% | 29.6% | Muse Wonder |
| CUA-bench (Vals AI) | 7.2% | Not measured | Only Muse Wonder measured |
| Vending-Bench 2 | $9,432 | Not measured | Only Muse Wonder measured |
| Neither model has a result on the other 26 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| LiveBench reasoning | Not measured | 83.8% | Only Step 5 Preview measured |
| MysteryMechanism (Vals AI) | 27.5% | 18.5% | Muse Wonder |
| Humanity's Last Exam (Artificial Analysis run) | 50.1% | 46.5% | Muse Wonder |
| LMArena Text: hard prompts | 1506 | 1461 | Muse Wonder |
| LMArena Text: expert | 1513 | 1485 | Muse Wonder |
| LMArena Text: hard prompts in English | 1506 | 1454 | Muse Wonder |
| LMArena Vision: diagrams | Not measured | 1308 | Only Step 5 Preview measured |
| LMArena Vision: homework | Not measured | 1265 | Only Step 5 Preview measured |
| Neither model has a result on the other 14 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| ProofBench | 84.0% | 42.0% | Muse Wonder |
| LMArena Text: math | 1486 | 1458 | Muse Wonder |
| LMArena Text: mathematical work | 1514 | 1460 | Muse Wonder |
| Neither model has a result on the other 10 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| LMArena Text: science | 1506 | 1454 | Muse Wonder |
| LMArena Text: medicine and healthcare | 1512 | 1439 | Muse Wonder |
| LMArena Text: legal and government | 1505 | 1437 | Muse Wonder |
| LMArena Text: business and finance | 1483 | 1435 | Muse Wonder |
| Neither model has a result on the other 14 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1362 | Not measured | Only Muse Wonder measured |
| Design Arena: UI components | 1374 | Not measured | Only Muse Wonder measured |
| Design Arena: game dev | 1377 | Not measured | Only Muse Wonder measured |
| Design Arena: data viz | 1321 | Not measured | Only Muse Wonder measured |
| Design Arena: 3D | 1426 | Not measured | Only Muse Wonder measured |
| Design Arena: all code categories | 1371 | Not measured | Only Muse Wonder measured |
| LiveBench design | Not measured | 76.9% | Only Step 5 Preview measured |
| Neither model has a result on the other 2 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index131
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| Vals Index | 64.9% | Not measured | Only Muse Wonder measured |
| LiveBench data analysis | Not measured | 63.5% | Only Step 5 Preview measured |
| Public Benefits Bench (Vals AI) | 70.2% | 58.1% | Muse Wonder |
| Long context reasoning (AA-LCR) | 80.7% | 88.3% | Step 5 Preview |
| DTBench | 95.2% | Not measured | Only Muse Wonder measured |
| LMCA | 56.8% | Not measured | Only Muse Wonder measured |
| LMArena Text | 1502 | 1440 | Muse Wonder |
| LMArena Text: instruction following | 1491 | 1438 | Muse Wonder |
| LMArena Text: creative writing | 1491 | 1409 | Muse Wonder |
| LMArena Vision | Not measured | 1265 | Only Step 5 Preview measured |
| LMArena Text: English | 1507 | 1447 | Muse Wonder |
| LMArena Text: non-English | 1492 | 1428 | Muse Wonder |
| LMArena Text: Chinese | 1537 | 1500 | Muse Wonder |
| LMArena Text: Japanese | Not measured | 1383 | Only Step 5 Preview measured |
| LMArena Text: French | 1523 | 1438 | Muse Wonder |
| LMArena Text: Spanish | Not measured | 1441 | Only Step 5 Preview measured |
| LMArena Text: Russian | 1522 | 1433 | Muse Wonder |
| LMArena Text: multi-turn | 1510 | 1437 | Muse Wonder |
| LMArena Text: longer queries | 1493 | 1449 | Muse Wonder |
| LMArena Text: writing and language | 1496 | 1425 | Muse Wonder |
| LMArena Text: entertainment and media | 1482 | 1412 | Muse Wonder |
| LMArena Vision: English | Not measured | 1290 | Only Step 5 Preview measured |
| LMArena Vision: reading text | Not measured | 1282 | Only Step 5 Preview measured |
| Neither model has a result on the other 23 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Muse Wonder | Step 5 Preview | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Step 5 Preview costs less
Repository review
50K input + 3K output tokens
Step 5 Preview costs less
Agent loop
200K input + 10K output tokens
Step 5 Preview costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Meta Muse Wonder | StepFun Step 5 Preview |
|---|---|---|
| Maker | Meta | StepFun |
| Weights | No public weights listed | Open weights |
| Released | October 12, 2026 | October 8, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $4.00 | $1.00 |
| Output $/Mtok | $20.00 | $2.70 |
| Coding index | 71.9 | 61.1 |
| Independent measurements behind it | 3 | 4 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Muse Wonder better than Step 5 Preview for coding?
Pick Muse Wonder for coding: it scores 71.9 on the Vibe Coding Index against 61.1 for Step 5 Preview, though on 3 benchmarks against 4. Pick Step 5 Preview for output tokens 7.4x cheaper ($2.70 vs $20.00 per million) and work measured by BioMysteryBench (Vals AI) or Long context reasoning (AA-LCR), where it leads.
Which is cheaper, Muse Wonder or Step 5 Preview?
Muse Wonder costs $4.00 per million input tokens and $20.00 per million output tokens; Step 5 Preview costs $1.00 and $2.70. Cost per solved task is known for neither of them.
Which has the bigger context window, Muse Wonder or Step 5 Preview?
Neither: both take 1.05M tokens.
How many benchmarks is the Muse Wonder vs Step 5 Preview comparison based on?
44 benchmarks have a result for both models as of October 21, 2026. Muse Wonder's coding index rests on 3 independent benchmarks and Step 5 Preview's on 4. Where only one model has a result, the page says so instead of naming a winner.