Model comparison
Gemma 4 31B IT vs GPT-6 Luna
GPT-6 Luna is the stronger model overall, scoring 53.3 to 43.5 on the Noometry Index.
Last verified . 27 shared benchmarks.
GoogleSummary
- They share 27 benchmarks with published results for both. Gemma 4 31B IT scores higher in 4 categories and GPT-6 Luna in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 43.2.
- The biggest single-benchmark swing is SimpleQA Verified: 10.4% for Gemma 4 31B IT and 41.4% for GPT-6 Luna.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $0.10 / $0.50 for GPT-6 Luna.
- GPT-6 Luna accepts more context: 1.05M tokens versus 262K.
- Gemma 4 31B IT has downloadable open weights; the other is API-only.
Side by side
| Gemma 4 31B IT | GPT-6 Luna | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 43.5 | 53.3 |
| Released | 2026-04-02 | 2026-09-22 |
| Weights | Open | Proprietary |
| Context window | 262K | 1.05M |
| Max output | 33K | 128K |
| Input $ / M tokens | $0.09 | $0.10 |
| Output $ / M tokens | $0.34 | $0.50 |
| Results tracked | 35 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
Gemma 4 31B IT: 42.3 (#108), GPT-6 Luna: 55.5 (#25)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena WebDev | 1366 | 1581 |
| SciCode | 43.4% | 54.6% |
| LMArena Coding | 1459 | 1439 |
| ALE-Bench | 925.5 | 1,577 |
| DeepSWE | — | 66.6% |
| FrontierCode | — | 42.4% |
| WeirdML | 52.3% | — |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, GPT-6 Luna: 33.3 (#54)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| APEX-Agents | — | 44.3% |
| GDP.pdf | — | 23% |
Reasoning GPT-6 Luna leads
Gemma 4 31B IT: 27.2 (#122), GPT-6 Luna: 48.2 (#41)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| NYT Connections (extended) | 70.6% | 68.7% |
| CritPt | 1.4% | 19.4% |
| Chess Puzzles | 5% | 31% |
| LMArena Hard Prompts | 1448 | 1411 |
| DTBench | 82.7% | 90.1% |
| LMCA | 39.3% | 44.5% |
| Epoch Capabilities Index | 142.74 | 156.28 |
| ARC-AGI-2 | — | 59.3% |
| Kagi LLM Benchmark | 63.5% | — |
| ARC-AGI-1 | — | 86.7% |
| Thematic Generalization | 53% | — |
| Mystery Game Puzzles | — | 7% |
| Surface Evolver Bench | 30.6% | — |
Math GPT-6 Luna leads
Gemma 4 31B IT: 43.2 (#81), GPT-6 Luna: 76.1 (#15)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 98.9% |
| LMArena Math | 1465 | 1416 |
| FrontierMath (Tiers 1-3) | — | 78.9% |
| FrontierMath Tier 4 | — | 56.1% |
| ProofBench | — | 64% |
Knowledge GPT-6 Luna leads
Gemma 4 31B IT: 37.9 (#151), GPT-6 Luna: 57.0 (#41)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 75.8% | 90.5% |
| SimpleQA Verified | 10.4% | 41.4% |
| LMArena Expert | 1465 | 1444 |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Too close to call
Gemma 4 31B IT: 41.6 (#34), GPT-6 Luna: 42.4 (#30)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena Vision | 1277 | 1217 |
| Blueprint-Bench 2 | — | 31.2% |
| Furniture Assembly | — | 44.2% |
| LMArena Document | 1425 | — |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 53.8 (#57), GPT-6 Luna: 50.5 (#117)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1431 | 1386 |
| LMArena Chinese | 1476 | 1433 |
| LMArena French | 1435 | 1420 |
| LMArena Russian | 1460 | 1394 |
| LMArena Spanish | 1444 | 1393 |
| LMArena German | — | 1369 |
| LMArena Japanese | — | 1369 |
| LMArena Korean | — | 1360 |
Instruction Following Gemma 4 31B IT leads
Gemma 4 31B IT: 75.5 (#61), GPT-6 Luna: 74.3 (#99)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1433 | 1409 |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#71), GPT-6 Luna: 43.0 (#111)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1446 | 1409 |
Writing & Preference Gemma 4 31B IT leads
Gemma 4 31B IT: 60.5 (#96), GPT-6 Luna: 58.3 (#119)
| Benchmark | Gemma 4 31B IT | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1443 | 1391 |
| LMArena Creative Writing | 1415 | 1363 |
| LMArena Multi-Turn | 1452 | 1396 |
| EQ-Bench Creative Writing | 1368 | — |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than GPT-6 Luna?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 43.5 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or GPT-6 Luna?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; GPT-6 Luna lists at $0.10 and $0.50.
Is Gemma 4 31B IT or GPT-6 Luna better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 42.3 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 262K.
How many benchmarks do Gemma 4 31B IT and GPT-6 Luna share?
27 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and GPT-6 Luna has 42.