Gemma 3 4B vs Muse Wonder
Gemma 3 4B costs 266.7x less per output token than Muse Wonder ($0.075 vs $20.00 per million). Updated October 22, 2026; 26 benchmarks have a result for both.
Gemma 3 4B or Muse Wonder: which to pick
Only Muse Wonder has a coding score, 71.9 on the Vibe Coding Index; Gemma 3 4B has none, so the two cannot be ranked on coding.
Pick Gemma 3 4B for output tokens 266.7x cheaper ($0.075 vs $20.00 per million).
- Shared results
- 26
- Gemma 3 4B only
- 16
- Muse Wonder only
- 34
- 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
No call
Only Muse Wonder has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
No call
Only Gemma 3 4B has a AA Intelligence Index result, 4.8.
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
Muse Wonder
Muse Wonder takes 1.05M tokens against 131K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemma 3 4B
Gemma 3 4B $0.0001, Muse Wonder $0.014 per turn.
List prices
At a glance
- Coding index
- Gemma 3 4B not benchmarked, Muse Wonder 71.9 from 3 benchmarks.
- Widest category
- Intelligence, 57.5 points apart on average over 1 shared benchmark.
- Widest single gap
- Long context reasoning (AA-LCR): Gemma 3 4B 6.7% against Muse Wonder 80.7%, 74.0 points apart on a 0 to 100 scale.
- Output price
- Gemma 3 4B $0.075 against Muse Wonder $20.00 per million output tokens, 266.7x apart.
- Context window
- Gemma 3 4B takes 131K tokens, Muse Wonder 1.05M.
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 | Gemma 3 4B | Bars | Muse Wonder | Ties |
|---|---|---|---|---|
| Coding | 0 | 2 | ||
| Intelligence | 0 | 1 | ||
| Reasoning | 0 | 4 | ||
| Math | 0 | 2 | ||
| Knowledge | 0 | 4 | ||
| General | 0 | 13 | ||
| All shared boards | 0 | 26 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index02
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| LMArena WebDev | Not measured | 1652 | Only Muse Wonder measured |
| AA Coding Index | 2.7 | Not measured | Only Gemma 3 4B measured |
| Terminal-Bench 4 (Vals AI) | Not measured | 49.5% | Only Muse Wonder measured |
| Vibe Code Bench (Vals AI) | Not measured | 83.1% | Only Muse Wonder measured |
| Code migration (Vals AI) | Not measured | 62.5% | Only Muse Wonder measured |
| IOI (Vals AI) | Not measured | 66.9% | Only Muse Wonder measured |
| SWE-bench Multimodal (Vals AI) | Not measured | 45.6% | Only Muse Wonder measured |
| Vibe Code Bench 1-100 (Vals AI) | Not measured | 26.6% | Only Muse Wonder measured |
| CyberBench (Vals AI) | Not measured | 80.4% | Only Muse Wonder measured |
| LiveCodeBench (Artificial Analysis run) | 11.2% | Not measured | Only Gemma 3 4B measured |
| SciCode | Not measured | 60.3% | Only Muse Wonder measured |
| Terminal-Bench Hard (Artificial Analysis run) | 0.8% | Not measured | Only Gemma 3 4B measured |
| LMArena Text: coding | 1230 | 1522 | Muse Wonder |
| LMArena Text: software and IT | 1269 | 1515 | Muse Wonder |
| LMArena WebDev: React | Not measured | 1641 | Only Muse Wonder measured |
| LMArena WebDev: HTML | Not measured | 1694 | Only Muse Wonder measured |
| SciCode | 0.0% | Not measured | Only Gemma 3 4B measured |
| Neither model has a result on the other 22 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| AA Intelligence Index | 4.8 | Not measured | Only Gemma 3 4B measured |
| Epoch Capabilities Index | 115.9 | 161.8 | Muse Wonder |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index13 boards, none shared
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| Design Arena: spreadsheets | Not measured | 1243 | Only Muse Wonder measured |
| Finance Agent (Vals AI) | Not measured | 56.6% | Only Muse Wonder measured |
| Tax Agent Bench (Vals AI) | Not measured | 36.3% | Only Muse Wonder measured |
| Legal Research Bench (Vals AI) | Not measured | 50.5% | Only Muse Wonder measured |
| Harvey's Legal Agent Benchmark (Vals AI) | Not measured | 23.8% | Only Muse Wonder measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 67.1% | Only Muse Wonder measured |
| BioMysteryBench (Vals AI) | Not measured | 67.0% | Only Muse Wonder measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 22.9% | Only Muse Wonder measured |
| Vals Science Index | Not measured | 39.1% | Only Muse Wonder measured |
| CUA-bench (Vals AI) | Not measured | 7.2% | Only Muse Wonder measured |
| τ²-Bench Telecom | 5.0% | Not measured | Only Gemma 3 4B measured |
| Vending-Bench 2 | Not measured | $9,432 | Only Muse Wonder measured |
| Berkeley Function Calling Leaderboard (BFCL) | 19.6% | Not measured | Only Gemma 3 4B measured |
| Neither model has a result on the other 28 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| MysteryMechanism (Vals AI) | Not measured | 27.5% | Only Muse Wonder measured |
| GPQA Diamond (Artificial Analysis run) | 29.1% | Not measured | Only Gemma 3 4B measured |
| Humanity's Last Exam (Artificial Analysis run) | 5.3% | 50.1% | Muse Wonder |
| GPQA Diamond (Epoch AI run) | 23.2% | Not measured | Only Gemma 3 4B measured |
| LMArena Text: hard prompts | 1253 | 1506 | Muse Wonder |
| LMArena Text: expert | 1223 | 1513 | Muse Wonder |
| LMArena Text: hard prompts in English | 1253 | 1506 | Muse Wonder |
| Kagi LLM Benchmark | 25.2% | Not measured | Only Gemma 3 4B measured |
| Neither model has a result on the other 14 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| AA Math Index | 12.7 | Not measured | Only Gemma 3 4B measured |
| AIME 2025 (Artificial Analysis run) | 12.7% | Not measured | Only Gemma 3 4B measured |
| OTIS Mock AIME | 7.5% | Not measured | Only Gemma 3 4B measured |
| ProofBench | Not measured | 84.0% | Only Muse Wonder measured |
| LMArena Text: math | 1239 | 1486 | Muse Wonder |
| LMArena Text: mathematical work | 1256 | 1514 | Muse Wonder |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 41.7% | Not measured | Only Gemma 3 4B measured |
| LMArena Text: science | 1306 | 1506 | Muse Wonder |
| LMArena Text: medicine and healthcare | 1342 | 1512 | Muse Wonder |
| LMArena Text: legal and government | 1327 | 1505 | Muse Wonder |
| LMArena Text: business and finance | 1311 | 1483 | Muse Wonder |
| Neither model has a result on the other 13 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1362 | Only Muse Wonder measured |
| Design Arena: UI components | Not measured | 1379 | Only Muse Wonder measured |
| Design Arena: game dev | Not measured | 1376 | Only Muse Wonder measured |
| Design Arena: data viz | Not measured | 1317 | Only Muse Wonder measured |
| Design Arena: 3D | Not measured | 1427 | Only Muse Wonder measured |
| Design Arena: all code categories | Not measured | 1371 | Only Muse Wonder measured |
| Neither model has a result on the other 3 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index013
| Benchmark | Gemma 3 4B | Muse Wonder | Leader |
|---|---|---|---|
| Vals Index | Not measured | 64.9% | Only Muse Wonder measured |
| Public Benefits Bench (Vals AI) | Not measured | 70.2% | Only Muse Wonder measured |
| IFBench | 28.3% | Not measured | Only Gemma 3 4B measured |
| Long context reasoning (AA-LCR) | 6.7% | 80.7% | Muse Wonder |
| DTBench | 50.9% | 95.2% | Muse Wonder |
| LMCA | 2.8% | 56.8% | Muse Wonder |
| LMArena Text | 1291 | 1502 | Muse Wonder |
| LMArena Text: instruction following | 1239 | 1491 | Muse Wonder |
| LMArena Text: creative writing | 1270 | 1491 | Muse Wonder |
| LMArena Text: English | 1300 | 1507 | Muse Wonder |
| LMArena Text: non-English | 1273 | 1492 | Muse Wonder |
| LMArena Text: Chinese | Not measured | 1537 | Only Muse Wonder measured |
| LMArena Text: German | 1281 | Not measured | Only Gemma 3 4B measured |
| LMArena Text: French | Not measured | 1523 | Only Muse Wonder measured |
| LMArena Text: Russian | 1294 | 1522 | Muse Wonder |
| LMArena Text: multi-turn | 1255 | 1510 | Muse Wonder |
| LMArena Text: longer queries | 1273 | 1493 | Muse Wonder |
| LMArena Text: writing and language | 1288 | 1496 | Muse Wonder |
| LMArena Text: entertainment and media | 1233 | 1482 | Muse Wonder |
| Neither model has a result on the other 27 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemma 3 4B | Muse Wonder | 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
Gemma 3 4B costs less
Repository review
50K input + 3K output tokens
Gemma 3 4B costs less
Agent loop
200K input + 10K output tokens
Gemma 3 4B 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 | Google Gemma 3 4B | Meta Muse Wonder |
|---|---|---|
| Maker | Meta | |
| Weights | Open weights | No public weights listed |
| Released | March 13, 2025 | October 12, 2026 |
| Context window | 131K | 1.05M |
| Input $/Mtok | $0.045 | $4.00 |
| Output $/Mtok | $0.075 | $20.00 |
| Coding index | Not benchmarked | 71.9 |
| Independent measurements behind it | 0 | 3 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Gemma 3 4B better than Muse Wonder for coding?
Only Muse Wonder has a coding score, 71.9 on the Vibe Coding Index; Gemma 3 4B has none, so the two cannot be ranked on coding. Pick Gemma 3 4B for output tokens 266.7x cheaper ($0.075 vs $20.00 per million).
Which is cheaper, Gemma 3 4B or Muse Wonder?
Gemma 3 4B costs $0.045 per million input tokens and $0.075 per million output tokens; Muse Wonder costs $4.00 and $20.00. Cost per solved task is known for neither of them.
Which has the bigger context window, Gemma 3 4B or Muse Wonder?
Muse Wonder, with 1.05M tokens against 131K.
How many benchmarks is the Gemma 3 4B vs Muse Wonder comparison based on?
26 benchmarks have a result for both models as of October 22, 2026. Gemma 3 4B's coding index rests on 0 independent benchmarks and Muse Wonder's on 3. Where only one model has a result, the page says so instead of naming a winner.