Model comparison
Gemma 4 31B IT vs GPT-5 Nano
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 33.5 on the Noometry Index.
Last verified . 26 shared benchmarks.
GoogleSummary
- They share 26 benchmarks with published results for both. Gemma 4 31B IT scores higher in 9 categories and GPT-5 Nano in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 4 31B IT leads 60.5 to 39.1.
- The biggest single-benchmark swing is LMCA: 39.3% for Gemma 4 31B IT and 7.9% for GPT-5 Nano.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.09 / $0.34 for Gemma 4 31B IT.
- GPT-5 Nano 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 Nano | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 43.5 | 33.5 |
| Released | 2026-04-02 | 2025-08-07 |
| Weights | Open | Proprietary |
| Context window | 262K | 400K |
| Max output | 33K | 128K |
| Input $ / M tokens | $0.09 | $0.05 |
| Output $ / M tokens | $0.34 | $0.40 |
| Results tracked | 35 | 49 |
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 Nano: 33.6 (#254)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| WeirdML | 52.3% | 38.1% |
| LMArena Coding | 1459 | 1351 |
| ALE-Bench | 925.5 | 718.67 |
| SWE-bench Verified (bash only) | — | 34.8% |
| LMArena WebDev | 1366 | — |
| SciCode | 43.4% | — |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, GPT-5 Nano: 25.8 (#106)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| Terminal-Bench | — | 21.8% |
| Berkeley Function Calling Leaderboard | — | 51.5% |
Reasoning Gemma 4 31B IT leads
Gemma 4 31B IT: 27.2 (#122), GPT-5 Nano: 16.3 (#306)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 62.2% |
| Chess Puzzles | 5% | 27% |
| LMArena Hard Prompts | 1448 | 1328 |
| DTBench | 82.7% | 62.7% |
| LMCA | 39.3% | 7.9% |
| Epoch Capabilities Index | 142.74 | 139.38 |
| ARC-AGI-2 | — | 2.6% |
| NYT Connections (extended) | 70.6% | — |
| ARC-AGI-1 | — | 20.7% |
| CritPt | 1.4% | — |
| Thematic Generalization | 53% | — |
| Mystery Game Puzzles | — | 9% |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 59.1 |
Math Gemma 4 31B IT leads
Gemma 4 31B IT: 43.2 (#81), GPT-5 Nano: 29.4 (#241)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 81.1% |
| LMArena Math | 1465 | 1317 |
| FrontierMath (Tiers 1-3) | — | 20% |
| FrontierMath Tier 4 | — | 2.4% |
| ProofBench | — | 12% |
| Omni-MATH | — | 54.6% |
| MATH Level 5 | — | 95.2% |
| FrontierMath (Feb 2025 set) | — | 8.3% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge Gemma 4 31B IT leads
Gemma 4 31B IT: 37.9 (#151), GPT-5 Nano: 35.9 (#178)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| GPQA Diamond | 75.8% | 69.4% |
| SimpleQA Verified | 10.4% | 11.7% |
| Vectara Hallucination Rate | 7.4% | 10.5% |
| LMArena Expert | 1465 | 1321 |
| MMLU-Pro | — | 77.8% |
| GPQA (HELM) | — | 67.9% |
Multimodal Gemma 4 31B IT leads
Gemma 4 31B IT: 41.6 (#34), GPT-5 Nano: 31.3 (#108)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| LMArena Vision | 1277 | 1159 |
| VPCT | — | 37.2% |
| LMArena Document | 1425 | — |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 53.8 (#57), GPT-5 Nano: 45.3 (#172)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| LMArena Non-English | 1431 | 1313 |
| LMArena Chinese | 1476 | 1356 |
| LMArena Russian | 1460 | 1296 |
| LMArena Spanish | 1444 | 1360 |
| LMArena French | 1435 | — |
| LMArena German | — | 1327 |
| LMArena Japanese | — | 1226 |
| LMArena Korean | — | 1269 |
Instruction Following Too close to call
Gemma 4 31B IT: 75.5 (#61), GPT-5 Nano: 75.0 (#79)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| LMArena Instruction Following | 1433 | 1306 |
| IFEval | — | 93.2% |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#71), GPT-5 Nano: 31.3 (#281)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| LMArena Longer Query | 1446 | 1312 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference Gemma 4 31B IT leads
Gemma 4 31B IT: 60.5 (#96), GPT-5 Nano: 39.1 (#249)
| Benchmark | Gemma 4 31B IT | GPT-5 Nano |
|---|---|---|
| LMArena Text | 1443 | 1320 |
| LMArena Creative Writing | 1415 | 1249 |
| EQ-Bench Creative Writing | 1368 | 705 |
| LMArena Multi-Turn | 1452 | 1311 |
| WildBench | — | 80.6% |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than GPT-5 Nano?
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 33.5 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or GPT-5 Nano?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Gemma 4 31B IT lists at $0.09 and $0.34.
Is Gemma 4 31B IT or GPT-5 Nano better for coding?
Gemma 4 31B IT scores higher on coding benchmarks: 42.3 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 262K.
How many benchmarks do Gemma 4 31B IT and GPT-5 Nano share?
26 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and GPT-5 Nano has 49.