Model comparison
Gemma 3 4B vs gpt-oss-120b
gpt-oss-120b is the stronger model overall, scoring 36.3 to 28.1 on the Noometry Index.
Last verified . 21 shared benchmarks.
GoogleSummary
- They share 21 benchmarks with published results for both. Gemma 3 4B scores higher in 3 categories and gpt-oss-120b in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 52.5 to 16.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 88.9% for gpt-oss-120b.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.037 / $0.17 for gpt-oss-120b.
Side by side
| Gemma 3 4B | gpt-oss-120b | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 28.1 | 36.3 |
| Released | 2025-03-12 | 2025-08-05 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 4K | 41K |
| Input $ / M tokens | $0.04 | $0.037 |
| Output $ / M tokens | $0.08 | $0.17 |
| Results tracked | 22 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 3 4B leads
Gemma 3 4B: 35.9 (#215), gpt-oss-120b: 33.5 (#256)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| LMArena Coding | 1230 | 1380 |
| SWE-bench Verified (bash only) | — | 26% |
| Aider Polyglot | — | 41.8% |
| SciCode | — | 36% |
| WeirdML | — | 48.2% |
| ALE-Bench | — | 575.62 |
| AlgoTune | — | 1.41 |
Agentic & Tool Use Gemma 3 4B leads
Gemma 3 4B: 20.9 (#142), gpt-oss-120b: 12.2 (#153)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| Berkeley Function Calling Leaderboard | 19.6% | — |
| METR Time Horizons | — | 56.6% |
| Vending-Bench 2 | — | -21.53 |
Reasoning gpt-oss-120b leads
Gemma 3 4B: 13.2 (#335), gpt-oss-120b: 20.0 (#245)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| Kagi LLM Benchmark | 25.2% | 58.6% |
| Chess Puzzles | 0% | 20% |
| LMArena Hard Prompts | 1253 | 1364 |
| DTBench | 50.9% | 76.3% |
| LMCA | 2.8% | 22.1% |
| Epoch Capabilities Index | 116.02 | 139.93 |
| SimpleBench | — | 22.1% |
| CritPt | — | 1.1% |
| Mystery Game Puzzles | — | 2% |
| Surface Evolver Bench | — | 25% |
Math gpt-oss-120b leads
Gemma 3 4B: 16.8 (#292), gpt-oss-120b: 52.5 (#50)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 88.9% |
| LMArena Math | 1239 | 1389 |
| Omni-MATH | — | 68.8% |
Knowledge gpt-oss-120b leads
Gemma 3 4B: 11.8 (#299), gpt-oss-120b: 42.4 (#96)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 23.2% | 75.8% |
| Vectara Hallucination Rate | 6.4% | 14.2% |
| LMArena Expert | 1223 | 1356 |
| MMLU-Pro | — | 79.5% |
| Confabulations | — | 15.7% |
| GPQA (HELM) | — | 68.4% |
Multilingual gpt-oss-120b leads
Gemma 3 4B: 42.5 (#194), gpt-oss-120b: 48.0 (#147)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1273 | 1351 |
| LMArena German | 1281 | 1353 |
| LMArena Russian | 1294 | 1343 |
| LMArena Chinese | — | 1385 |
| LMArena French | — | 1369 |
| LMArena Japanese | — | 1331 |
| LMArena Korean | — | 1282 |
| LMArena Spanish | — | 1389 |
Instruction Following gpt-oss-120b leads
Gemma 3 4B: 65.2 (#225), gpt-oss-120b: 69.3 (#173)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| LMArena Instruction Following | 1239 | 1318 |
| IFEval | — | 83.6% |
Long Context Gemma 3 4B leads
Gemma 3 4B: 38.7 (#194), gpt-oss-120b: 31.4 (#278)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| LMArena Longer Query | 1273 | 1319 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference gpt-oss-120b leads
Gemma 3 4B: 42.0 (#239), gpt-oss-120b: 46.5 (#217)
| Benchmark | Gemma 3 4B | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1291 | 1365 |
| LMArena Creative Writing | 1271 | 1275 |
| EQ-Bench Creative Writing | 1068 | 961 |
| LMArena Multi-Turn | 1255 | 1340 |
| Short-Story Creative Writing | — | 77.1% |
| WildBench | — | 84.5% |
Frequently asked questions
Is Gemma 3 4B better than gpt-oss-120b?
gpt-oss-120b is the stronger model overall, scoring 36.3 to 28.1 on the Noometry Index.
Which is cheaper, Gemma 3 4B or gpt-oss-120b?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; gpt-oss-120b lists at $0.037 and $0.17.
Is Gemma 3 4B or gpt-oss-120b better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 131K tokens.
How many benchmarks do Gemma 3 4B and gpt-oss-120b share?
21 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and gpt-oss-120b has 48.