Model comparison
Gemma 4 31B IT vs Llama 4 Maverick
Gemma 4 31B IT is the stronger model overall, scoring 43.2 to 30.6 on the Noometry Index.
Last verified . 28 shared benchmarks.
Google
MetaSummary
- They share 28 benchmarks with published results for both. Gemma 4 31B IT scores higher in 9 categories and Llama 4 Maverick in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 4 31B IT leads 60.7 to 39.0.
- The biggest single-benchmark swing is NYT Connections (extended): 70.6% for Gemma 4 31B IT and 8% for Llama 4 Maverick.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $0.19 / $0.65 for Llama 4 Maverick.
- Gemma 4 31B IT accepts more context: 262K tokens versus 128K.
Side by side
| Gemma 4 31B IT | Llama 4 Maverick | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 43.2 | 30.6 |
| Released | 2026-04-02 | 2025-04-05 |
| Weights | Open | Open |
| Context window | 262K | 128K |
| Max output | 33K | 4K |
| Input $ / M tokens | $0.09 | $0.19 |
| Output $ / M tokens | $0.34 | $0.65 |
| Results tracked | 35 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 4 31B IT leads
Gemma 4 31B IT: 41.9 (#98), Llama 4 Maverick: 26.3 (#318)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| SciCode | 43.4% | 33.1% |
| WeirdML | 52.3% | 24.5% |
| LMArena Coding | 1459 | 1302 |
| ALE-Bench | 925.5 | 172.97 |
| SWE-bench Verified (bash only) | — | 21% |
| Aider Polyglot | — | 15.6% |
| LMArena WebDev | 1364 | — |
| BigCodeBench Instruct | — | 49.7% |
| BigCodeBench Complete | — | 61.4% |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, Llama 4 Maverick: 27.9 (#86)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.3% |
Reasoning Gemma 4 31B IT leads
Gemma 4 31B IT: 26.4 (#119), Llama 4 Maverick: 10.1 (#337)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 55.9% |
| NYT Connections (extended) | 70.6% | 8% |
| CritPt | 1.4% | 0% |
| LMArena Hard Prompts | 1448 | 1281 |
| DTBench | 82.7% | 61.9% |
| LMCA | 39.3% | 15.9% |
| Epoch Capabilities Index | 142.74 | 132.2 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 27.7% |
| ARC-AGI-1 | — | 4.4% |
| Chess Puzzles | 5% | — |
| EnigmaEval | — | 0.6% |
| Thematic Generalization | 53% | — |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 57.5 |
Math Gemma 4 31B IT leads
Gemma 4 31B IT: 41.3 (#72), Llama 4 Maverick: 24.6 (#256)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 20.6% |
| LMArena Math | 1464 | 1299 |
| Omni-MATH | — | 42.2% |
| MATH Level 5 | — | 73% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Gemma 4 31B IT leads
Gemma 4 31B IT: 37.9 (#141), Llama 4 Maverick: 33.3 (#196)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 75.8% | 67% |
| Vectara Hallucination Rate | 7.4% | 8.2% |
| LMArena Expert | 1465 | 1259 |
| Humanity's Last Exam | — | 5.7% |
| SimpleQA Verified | 10.4% | — |
| MMLU-Pro | — | 81% |
| Confabulations | — | 22.6% |
| GPQA (HELM) | — | 65% |
Multimodal Gemma 4 31B IT leads
Gemma 4 31B IT: 41.0 (#32), Llama 4 Maverick: 30.9 (#100)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | 1278 | 1141 |
| GeoBench | — | 52% |
| LMArena Document | 1425 | — |
| SpatialViz-Bench | — | 31.8% |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 54.0 (#52), Llama 4 Maverick: 42.4 (#190)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1431 | 1269 |
| LMArena Chinese | 1478 | 1277 |
| LMArena French | 1436 | 1258 |
| LMArena Russian | 1462 | 1286 |
| LMArena Spanish | 1443 | 1293 |
| LMArena German | — | 1291 |
| LMArena Japanese | — | 1207 |
| LMArena Korean | — | 1203 |
Instruction Following Gemma 4 31B IT leads
Gemma 4 31B IT: 75.6 (#58), Llama 4 Maverick: 71.9 (#140)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| LMArena Instruction Following | 1433 | 1267 |
| IFEval | — | 90.8% |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#67), Llama 4 Maverick: 31.4 (#274)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| LMArena Longer Query | 1446 | 1280 |
| Fiction.LiveBench | — | 46.2% |
Writing & Preference Gemma 4 31B IT leads
Gemma 4 31B IT: 60.7 (#92), Llama 4 Maverick: 39.0 (#247)
| Benchmark | Gemma 4 31B IT | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1443 | 1287 |
| LMArena Creative Writing | 1416 | 1266 |
| EQ-Bench Creative Writing | 1368 | 860 |
| LMArena Multi-Turn | 1452 | 1289 |
| Short-Story Creative Writing | — | 62% |
| WildBench | — | 80% |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than Llama 4 Maverick?
Gemma 4 31B IT is the stronger model overall, scoring 43.2 to 30.6 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or Llama 4 Maverick?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; Llama 4 Maverick lists at $0.19 and $0.65.
Is Gemma 4 31B IT or Llama 4 Maverick better for coding?
Gemma 4 31B IT scores higher on coding benchmarks: 41.9 versus 26.3 in the Noometry coding category.
Which has the bigger context window?
Gemma 4 31B IT does, with 262K tokens against 128K.
How many benchmarks do Gemma 4 31B IT and Llama 4 Maverick share?
28 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Llama 4 Maverick has 54.