Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
Gemma 4 31B
73
Winner · 1/8 categoriesGrok 4.1 Fast
70
2/8 categoriesGemma 4 31B· Grok 4.1 Fast
Pick Gemma 4 31B if you want the stronger benchmark profile. Grok 4.1 Fast only becomes the better choice if reasoning is the priority or you need the larger 1M context window.
Gemma 4 31B has the cleaner overall profile here, landing at 73 versus 70. It is a real lead, but still close enough that category-level strengths matter more than the headline number.
Gemma 4 31B's sharpest advantage is in coding, where it averages 80 against 60.7. The single biggest benchmark swing on the page is LiveCodeBench, 80% to 54%. Grok 4.1 Fast does hit back in reasoning, so the answer changes if that is the part of the workload you care about most.
Gemma 4 31B is the reasoning model in the pair, while Grok 4.1 Fast is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. Grok 4.1 Fast gives you the larger context window at 1M, compared with 256K for Gemma 4 31B.
BenchLM keeps the benchmark table and the operator tradeoffs on the same page so a better score does not hide a materially slower, pricier, or smaller-context model.
Runtime metrics show N/A when BenchLM does not have a sourced snapshot for that exact model. The scoring rules and freshness policy are documented on the methodology page.
| Benchmark | Gemma 4 31B | Grok 4.1 Fast |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 74% |
| BrowseComp | — | 73% |
| OSWorld-Verified | — | 66% |
| Claw-Eval | — | 53.8% |
| DeepPlanning | — | 19.1% |
| CodingGemma 4 31B wins | ||
| LiveCodeBench | 80% | 54% |
| HumanEval | — | 86% |
| SWE-bench Verified | — | 68% |
| SWE-bench Pro | — | 63% |
| Multimodal & Grounded | ||
| MMMU-Pro | 76.9% | — |
| ReasoningGrok 4.1 Fast wins | ||
| BBH | 74.4% | 87% |
| MRCRv2 | 66.4% | 89% |
| MuSR | — | 88% |
| LongBench v2 | — | 87% |
| KnowledgeGrok 4.1 Fast wins | ||
| GPQA | 84.3% | 92% |
| MMLU-Pro | 85.2% | 81% |
| HLE | 26.5% | 20% |
| HLE w/o tools | 19.5% | — |
| MMLU | — | 94% |
| SuperGPQA | — | 90% |
| FrontierScience | — | 83% |
| SimpleQA | — | 90% |
| Instruction Following | ||
| IFEval | — | 90% |
| Multilingual | ||
| MGSM | — | 88% |
| MMLU-ProX | — | 83% |
| Mathematics | ||
| AIME 2023 | — | 96% |
| AIME 2024 | — | 98% |
| AIME 2025 | — | 97% |
| HMMT Feb 2023 | — | 92% |
| HMMT Feb 2024 | — | 94% |
| HMMT Feb 2025 | — | 93% |
| BRUMO 2025 | — | 95% |
| MATH-500 | — | 89% |
Gemma 4 31B is ahead overall, 73 to 70. The biggest single separator in this matchup is LiveCodeBench, where the scores are 80% and 54%.
Grok 4.1 Fast has the edge for knowledge tasks in this comparison, averaging 70.9 versus 61.3. Inside this category, GPQA is the benchmark that creates the most daylight between them.
Gemma 4 31B has the edge for coding in this comparison, averaging 80 versus 60.7. Inside this category, LiveCodeBench is the benchmark that creates the most daylight between them.
Grok 4.1 Fast has the edge for reasoning in this comparison, averaging 87.9 versus 66.4. Inside this category, MRCRv2 is the benchmark that creates the most daylight between them.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.