Gemma 4 31B vs Muse Spark 1.1
Muse Spark 1.1 leads Gemma 4 31B on coding, 70.3 to 39.0 on the Vibe Coding Index, and Gemma 4 31B costs 10.6x less per output token ($0.40 vs $4.25 per million). Updated October 22, 2026; 51 benchmarks have a result for both.
Gemma 4 31B or Muse Spark 1.1: which to pick
Pick Muse Spark 1.1 for coding: it scores 70.3 on the Vibe Coding Index against 39.0 for Gemma 4 31B.
Pick Gemma 4 31B for output tokens 10.6x cheaper ($0.40 vs $4.25 per million) and work measured by LMArena Vision: homework or SAGE (Vals AI), where it leads.
- Shared results
- 51
- Gemma 4 31B only
- 20
- Muse Spark 1.1 only
- 45
- Like-for-like categories
- 7 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 Spark 1.1
Muse Spark 1.1 scores 70.3 against 39.0.
2 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
Muse Spark 1.1
Muse Spark 1.1 leads, 33.7 against 14.7.
Same board, both measured
Agentic work
AA Agentic Index
Muse Spark 1.1
Muse Spark 1.1 leads, 25.8 against 4.2.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only Muse Spark 1.1 has a GPQA Diamond result, 91.2%.
Directional only
Knowledge
MMLU Pro
No call
Only Muse Spark 1.1 has a MMLU Pro result, 88.7%.
Directional only
Web design
Design Arena: websites
No call
Only Muse Spark 1.1 has a Design Arena: websites result, 1270.
Directional only
Long documents
Listed context window
Muse Spark 1.1
Muse Spark 1.1 takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemma 4 31B
Gemma 4 31B $0.0003, Muse Spark 1.1 $0.0034 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Muse Spark 1.1 has the number, $0.43.
Directional only
At a glance
- Coding index
- Gemma 4 31B 39.0 from 2 benchmarks, Muse Spark 1.1 70.3 from 3 benchmarks.
- Widest category
- Agentic work, 21.6 points apart on average over 1 shared benchmark.
- Widest single gap
- SimpleQA Verified: Gemma 4 31B 10.4% against Muse Spark 1.1 57.8%, 47.4 points apart on a 0 to 100 scale.
- Output price
- Gemma 4 31B $0.40 against Muse Spark 1.1 $4.25 per million output tokens, 10.6x apart.
- Context window
- Gemma 4 31B takes 262K tokens, Muse Spark 1.1 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 4 31B | Bars | Muse Spark 1.1 | Ties |
|---|---|---|---|---|
| Coding | 0 | 7 | ||
| Intelligence | 0 | 2 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 1 | 6 | ||
| Math | 0 | 2 | ||
| Knowledge | 0 | 6 | ||
| General | 2 | 24 | ||
| All shared boards | 3 | 48 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index07
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| LiveBench coding | 49.0% | Not measured | Only Gemma 4 31B measured |
| LMArena WebDev | 1367 | 1542 | Muse Spark 1.1 |
| AA Coding Index | 43.4 | 71.3 | Muse Spark 1.1 |
| LiveCodeBench (Vals AI) | Not measured | 85.9% | Only Muse Spark 1.1 measured |
| SWE-bench Verified (Vals AI) | Not measured | 82.0% | Only Muse Spark 1.1 measured |
| Vibe Code Bench (Vals AI) | Not measured | 72.2% | Only Muse Spark 1.1 measured |
| Code migration (Vals AI) | Not measured | 31.1% | Only Muse Spark 1.1 measured |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Muse Spark 1.1 measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 69.3% | Only Muse Spark 1.1 measured |
| SkillsBench (Vals AI) | Not measured | 59.2% | Only Muse Spark 1.1 measured |
| SciCode | 45.5% | 58.8% | Muse Spark 1.1 |
| Terminal-Bench Hard (Artificial Analysis run) | 36.4% | Not measured | Only Gemma 4 31B measured |
| WeirdML | 52.3% | Not measured | Only Gemma 4 31B measured |
| ALE-Bench | 925.5 | Not measured | Only Gemma 4 31B measured |
| LMArena Text: coding | 1459 | 1498 | Muse Spark 1.1 |
| LMArena Text: software and IT | 1463 | 1498 | Muse Spark 1.1 |
| LMArena WebDev: React | 1359 | 1532 | Muse Spark 1.1 |
| LMArena WebDev: HTML | 1380 | 1540 | Muse Spark 1.1 |
| LMArena WebDev: from an image | Not measured | 1540 | Only Muse Spark 1.1 measured |
| SciCode | 8.9% | Not measured | Only Gemma 4 31B measured |
| Neither model has a result on the other 19 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| AA Intelligence Index | 14.7 | 33.7 | Muse Spark 1.1 |
| Epoch Capabilities Index | 143.0 | 154.1 | Muse Spark 1.1 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| AA Agentic Index | 4.2 | 25.8 | Muse Spark 1.1 |
| Design Arena: agents building web apps | Not measured | 1207 | Only Muse Spark 1.1 measured |
| Design Arena: agents building full-stack apps | Not measured | 1199 | Only Muse Spark 1.1 measured |
| Design Arena: agents building mobile apps | Not measured | 1175 | Only Muse Spark 1.1 measured |
| Design Arena: agents building Android apps | Not measured | 1178 | Only Muse Spark 1.1 measured |
| Design Arena: agents building Godot games | Not measured | 1103 | Only Muse Spark 1.1 measured |
| Design Arena: agents building games | Not measured | 1183 | Only Muse Spark 1.1 measured |
| Design Arena: HTML slides | Not measured | 1145 | Only Muse Spark 1.1 measured |
| Design Arena: PowerPoint slides | Not measured | 1134 | Only Muse Spark 1.1 measured |
| Design Arena: spreadsheets | Not measured | 1021 | Only Muse Spark 1.1 measured |
| LiveBench research | 44.4% | Not measured | Only Gemma 4 31B measured |
| LiveBench browser | 4.0% | Not measured | Only Gemma 4 31B measured |
| LiveBench long agentic | 13.3% | Not measured | Only Gemma 4 31B measured |
| Finance Agent (Vals AI) | Not measured | 57.2% | Only Muse Spark 1.1 measured |
| Legal Research Bench (Vals AI) | Not measured | 38.0% | Only Muse Spark 1.1 measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 56.4% | Only Muse Spark 1.1 measured |
| Time Horizon Index (Vals AI) | Not measured | 1.7% | Only Muse Spark 1.1 measured |
| τ²-Bench Telecom | 65.5% | Not measured | Only Gemma 4 31B measured |
| APEX-Agents | Not measured | 31.8% | Only Muse Spark 1.1 measured |
| Vending-Bench 2 | Not measured | $6,520 | Only Muse Spark 1.1 measured |
| Neither model has a result on the other 21 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index16
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 91.2% | Only Muse Spark 1.1 measured |
| LiveBench reasoning | 68.3% | Not measured | Only Gemma 4 31B measured |
| MMMU Pro (Vals AI) | Not measured | 86.6% | Only Muse Spark 1.1 measured |
| GPQA Diamond (Artificial Analysis run) | 85.7% | 89.8% | Muse Spark 1.1 |
| Humanity's Last Exam (Artificial Analysis run) | 23.6% | 46.2% | Muse Spark 1.1 |
| GPQA Diamond (Epoch AI run) | 75.8% | Not measured | Only Gemma 4 31B measured |
| Chess puzzles | Not measured | 34.0% | Only Muse Spark 1.1 measured |
| LMArena Text: hard prompts | 1448 | 1485 | Muse Spark 1.1 |
| LMArena Text: expert | 1464 | 1478 | Muse Spark 1.1 |
| LMArena Text: hard prompts in English | 1456 | 1480 | Muse Spark 1.1 |
| LMArena Vision: diagrams | 1283 | 1305 | Muse Spark 1.1 |
| LMArena Vision: homework | 1310 | 1287 | Gemma 4 31B |
| Kagi LLM Benchmark | 63.5% | Not measured | Only Gemma 4 31B measured |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| OTIS Mock AIME | 73.3% | Not measured | Only Gemma 4 31B measured |
| ProofBench | Not measured | 39.0% | Only Muse Spark 1.1 measured |
| LMArena Text: math | 1463 | 1482 | Muse Spark 1.1 |
| LMArena Text: mathematical work | 1465 | 1487 | Muse Spark 1.1 |
| Neither model has a result on the other 9 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index06
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 88.7% | Only Muse Spark 1.1 measured |
| LegalBench (Vals AI) | Not measured | 85.0% | Only Muse Spark 1.1 measured |
| CaseLaw (Vals AI) | 52.6% | Not measured | Only Gemma 4 31B measured |
| CorpFin (Vals AI) | Not measured | 71.3% | Only Muse Spark 1.1 measured |
| TaxEval (Vals AI) | Not measured | 79.7% | Only Muse Spark 1.1 measured |
| MortgageTax (Vals AI) | 61.4% | 66.1% | Muse Spark 1.1 |
| SimpleQA Verified | 10.4% | 57.8% | Muse Spark 1.1 |
| LMArena Text: science | 1457 | 1484 | Muse Spark 1.1 |
| LMArena Text: medicine and healthcare | 1451 | 1479 | Muse Spark 1.1 |
| LMArena Text: legal and government | 1438 | 1481 | Muse Spark 1.1 |
| LMArena Text: business and finance | 1430 | 1470 | Muse Spark 1.1 |
| FACTS Benchmark Suite | 45.3% | Not measured | Only Gemma 4 31B measured |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1270 | Only Muse Spark 1.1 measured |
| Design Arena: UI components | Not measured | 1287 | Only Muse Spark 1.1 measured |
| Design Arena: game dev | Not measured | 1269 | Only Muse Spark 1.1 measured |
| Design Arena: data viz | Not measured | 1267 | Only Muse Spark 1.1 measured |
| Design Arena: 3D | Not measured | 1259 | Only Muse Spark 1.1 measured |
| Design Arena: all code categories | Not measured | 1276 | Only Muse Spark 1.1 measured |
| Design Arena: SVG | Not measured | 1234 | Only Muse Spark 1.1 measured |
| Design Arena: ASCII art | Not measured | 1269 | Only Muse Spark 1.1 measured |
| LiveBench design | 58.4% | Not measured | Only Gemma 4 31B measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index224
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | Leader |
|---|---|---|---|
| LiveBench data analysis | 53.3% | Not measured | Only Gemma 4 31B measured |
| Public Benefits Bench (Vals AI) | Not measured | 61.3% | Only Muse Spark 1.1 measured |
| SAGE (Vals AI) | 55.0% | 45.6% | Gemma 4 31B |
| MedScribe (Vals AI) | Not measured | 88.9% | Only Muse Spark 1.1 measured |
| Vals Multimodal Index | Not measured | 66.7% | Only Muse Spark 1.1 measured |
| IFBench | 75.6% | Not measured | Only Gemma 4 31B measured |
| Long context reasoning (AA-LCR) | 69.7% | 77.7% | Muse Spark 1.1 |
| DTBench | 82.7% | 94.4% | Muse Spark 1.1 |
| LMCA | 39.3% | 49.9% | Muse Spark 1.1 |
| Surface Evolver Bench | 30.6% | 52.5% | Muse Spark 1.1 |
| LMArena Text | 1443 | 1478 | Muse Spark 1.1 |
| LMArena Text: instruction following | 1433 | 1457 | Muse Spark 1.1 |
| LMArena Text: creative writing | 1417 | 1435 | Muse Spark 1.1 |
| LMArena Vision | 1276 | 1293 | Muse Spark 1.1 |
| LMArena Text: English | 1451 | 1478 | Muse Spark 1.1 |
| LMArena Text: non-English | 1430 | 1471 | Muse Spark 1.1 |
| LMArena Text: Chinese | 1475 | 1515 | Muse Spark 1.1 |
| LMArena Text: German | Not measured | 1469 | Only Muse Spark 1.1 measured |
| LMArena Text: Japanese | Not measured | 1450 | Only Muse Spark 1.1 measured |
| LMArena Text: Korean | Not measured | 1459 | Only Muse Spark 1.1 measured |
| LMArena Text: French | 1436 | 1492 | Muse Spark 1.1 |
| LMArena Text: Spanish | 1442 | 1464 | Muse Spark 1.1 |
| LMArena Text: Russian | 1460 | 1482 | Muse Spark 1.1 |
| LMArena Text: Polish | 1433 | 1499 | Muse Spark 1.1 |
| LMArena Text: multi-turn | 1452 | 1484 | Muse Spark 1.1 |
| LMArena Text: longer queries | 1446 | 1462 | Muse Spark 1.1 |
| LMArena Text: writing and language | 1432 | 1448 | Muse Spark 1.1 |
| LMArena Text: entertainment and media | 1405 | 1438 | Muse Spark 1.1 |
| LMArena Vision: English | 1271 | 1292 | Muse Spark 1.1 |
| LMArena Vision: Chinese | 1326 | 1348 | Muse Spark 1.1 |
| LMArena Vision: reading text | 1282 | 1301 | Muse Spark 1.1 |
| LMArena Vision: humor | 1280 | 1288 | Muse Spark 1.1 |
| LMArena Vision: captioning | 1218 | Not measured | Only Gemma 4 31B measured |
| LMArena Vision: recognizing things | 1226 | Not measured | Only Gemma 4 31B measured |
| LMArena Vision: creative writing | 1285 | 1253 | Gemma 4 31B |
| LMArena Document | 1425 | 1465 | Muse Spark 1.1 |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemma 4 31B | Muse Spark 1.1 | 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 4 31B costs less
Repository review
50K input + 3K output tokens
Gemma 4 31B costs less
Agent loop
200K input + 10K output tokens
Gemma 4 31B 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 4 31B | Meta Muse Spark 1.1 |
|---|---|---|
| Maker | Meta | |
| Weights | Open weights | No public weights listed |
| Released | April 2, 2026 | July 16, 2026 |
| Context window | 262K | 1.05M |
| Input $/Mtok | $0.14 | $1.25 |
| Output $/Mtok | $0.40 | $4.25 |
| Coding index | 39.0 | 70.3 |
| Independent measurements behind it | 2 | 3 |
| Cost per solved task | Not computable | $0.43 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Gemma 4 31B better than Muse Spark 1.1 for coding?
Pick Muse Spark 1.1 for coding: it scores 70.3 on the Vibe Coding Index against 39.0 for Gemma 4 31B. Pick Gemma 4 31B for output tokens 10.6x cheaper ($0.40 vs $4.25 per million) and work measured by LMArena Vision: homework or SAGE (Vals AI), where it leads.
Which is cheaper, Gemma 4 31B or Muse Spark 1.1?
Gemma 4 31B costs $0.14 per million input tokens and $0.40 per million output tokens; Muse Spark 1.1 costs $1.25 and $4.25. Cost per solved task is known for Muse Spark 1.1 only.
Which has the bigger context window, Gemma 4 31B or Muse Spark 1.1?
Muse Spark 1.1, with 1.05M tokens against 262K.
How many benchmarks is the Gemma 4 31B vs Muse Spark 1.1 comparison based on?
51 benchmarks have a result for both models as of October 22, 2026. Gemma 4 31B's coding index rests on 2 independent benchmarks and Muse Spark 1.1's on 3. Where only one model has a result, the page says so instead of naming a winner.