Model comparison
Gemma 4 31B IT vs GPT-5 Mini
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 41.8 on the Noometry Index.
Last verified . 29 shared benchmarks.
GoogleSummary
- They share 29 benchmarks with published results for both. Gemma 4 31B IT scores higher in 6 categories and GPT-5 Mini in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-5 Mini leads 45.6 to 37.9.
- The biggest single-benchmark swing is Chess Puzzles: 5% for Gemma 4 31B IT and 30% for GPT-5 Mini.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $0.25 / $2 for GPT-5 Mini.
- GPT-5 Mini accepts more context: 400K tokens versus 262K.
- Gemma 4 31B IT has downloadable open weights; the other is API-only.
Side by side
| Gemma 4 31B IT | GPT-5 Mini | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 43.5 | 41.8 |
| Released | 2026-04-02 | 2025-08-07 |
| Weights | Open | Proprietary |
| Context window | 262K | 400K |
| Max output | 33K | 128K |
| Input $ / M tokens | $0.09 | $0.25 |
| Output $ / M tokens | $0.34 | $2 |
| Results tracked | 35 | 60 |
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: 42.3 (#108), GPT-5 Mini: 40.1 (#146)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| SciCode | 43.4% | 39.2% |
| WeirdML | 52.3% | 52.7% |
| LMArena Coding | 1459 | 1406 |
| ALE-Bench | 925.5 | 799.77 |
| SWE-bench Verified | — | 64.7% |
| SWE-bench Verified (bash only) | — | 59.8% |
| LMArena WebDev | 1366 | — |
| SWE-bench Multilingual | — | 39.7% |
| AlgoTune | — | 1.38 |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, GPT-5 Mini: 31.1 (#70)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | — | 34.8% |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| Vending-Bench 2 | — | -31.18 |
Reasoning Gemma 4 31B IT leads
Gemma 4 31B IT: 27.2 (#122), GPT-5 Mini: 23.9 (#168)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 70.3% |
| CritPt | 1.4% | 0% |
| Chess Puzzles | 5% | 30% |
| LMArena Hard Prompts | 1448 | 1380 |
| DTBench | 82.7% | 80.5% |
| LMCA | 39.3% | 34.2% |
| Epoch Capabilities Index | 142.74 | 145.52 |
| ARC-AGI-2 | — | 4.4% |
| NYT Connections (extended) | 70.6% | — |
| ARC-AGI-1 | — | 54.3% |
| EnigmaEval | — | 8.2% |
| Thematic Generalization | 53% | — |
| Mystery Game Puzzles | — | 10% |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 61 |
Math GPT-5 Mini leads
Gemma 4 31B IT: 43.2 (#81), GPT-5 Mini: 46.7 (#69)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 86.7% |
| LMArena Math | 1465 | 1378 |
| FrontierMath (Tiers 1-3) | — | 46.7% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 9% |
| Omni-MATH | — | 72.2% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 27.2% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge GPT-5 Mini leads
Gemma 4 31B IT: 37.9 (#151), GPT-5 Mini: 45.6 (#86)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 75.8% | 75% |
| SimpleQA Verified | 10.4% | 21.6% |
| Vectara Hallucination Rate | 7.4% | 12.9% |
| LMArena Expert | 1465 | 1379 |
| Humanity's Last Exam | — | 19.4% |
| MMLU-Pro | — | 83.5% |
| Confabulations | — | 13.3% |
| GPQA (HELM) | — | 75.6% |
Multimodal Gemma 4 31B IT leads
Gemma 4 31B IT: 41.6 (#34), GPT-5 Mini: 35.6 (#85)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1277 | 1202 |
| VPCT | — | 40.2% |
| LMArena Document | 1425 | — |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 53.8 (#57), GPT-5 Mini: 48.9 (#137)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1431 | 1363 |
| LMArena Chinese | 1476 | 1385 |
| LMArena French | 1435 | 1386 |
| LMArena Russian | 1460 | 1362 |
| LMArena Spanish | 1444 | 1355 |
| LMArena German | — | 1366 |
| LMArena Japanese | — | 1341 |
| LMArena Korean | — | 1308 |
Instruction Following Too close to call
Gemma 4 31B IT: 75.5 (#61), GPT-5 Mini: 76.2 (#46)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1433 | 1357 |
| IFEval | — | 92.7% |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#71), GPT-5 Mini: 41.9 (#132)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1446 | 1355 |
| Fiction.LiveBench | — | 69.4% |
Writing & Preference Gemma 4 31B IT leads
Gemma 4 31B IT: 60.5 (#96), GPT-5 Mini: 55.2 (#148)
| Benchmark | Gemma 4 31B IT | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1443 | 1373 |
| LMArena Creative Writing | 1415 | 1325 |
| EQ-Bench Creative Writing | 1368 | 1313 |
| LMArena Multi-Turn | 1452 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| WildBench | — | 85.5% |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than GPT-5 Mini?
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 41.8 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or GPT-5 Mini?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; GPT-5 Mini lists at $0.25 and $2.
Is Gemma 4 31B IT or GPT-5 Mini better for coding?
Gemma 4 31B IT scores higher on coding benchmarks: 42.3 versus 40.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 262K.
How many benchmarks do Gemma 4 31B IT and GPT-5 Mini share?
29 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and GPT-5 Mini has 60.