Model comparison
Gemma 3 4B vs GPT-4.1
GPT-4.1 is the stronger model overall, scoring 35.7 to 27.9 on the Noometry Index. Gemma 3 4B costs 70× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
GoogleSummary
- They share 22 benchmarks with published results for both. Gemma 3 4B scores higher in 2 categories and GPT-4.1 in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-4.1 leads 37.0 to 11.8.
- The biggest single-benchmark swing is GPQA Diamond: 23.2% for Gemma 3 4B and 66.9% for GPT-4.1.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 4B | GPT-4.1 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 27.9 | 35.7 |
| Released | 2025-03-12 | 2025-04-14 |
| Weights | Open | Proprietary |
| Context window | 131K | 1.05M |
| Max output | 4K | 33K |
| Input $ / M tokens | $0.04 | $2 |
| Output $ / M tokens | $0.08 | $8 |
| Results tracked | 22 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 3 4B leads
Gemma 3 4B: 35.3 (#214), GPT-4.1: 34.2 (#228)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Coding | 1230 | 1391 |
| SWE-bench Verified | — | 48.5% |
| SWE-bench Verified (bash only) | — | 39.6% |
| Aider Polyglot | — | 52.4% |
| WeirdML | — | 39% |
| CadEval | — | 42% |
| ALE-Bench | — | 558.1 |
Agentic & Tool Use GPT-4.1 leads
Gemma 3 4B: 20.7 (#134), GPT-4.1: 34.4 (#40)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | 54% |
Reasoning Gemma 3 4B leads
Gemma 3 4B: 12.7 (#330), GPT-4.1: 11.7 (#334)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| Kagi LLM Benchmark | 25.2% | 52.3% |
| Chess Puzzles | 0% | 6% |
| LMArena Hard Prompts | 1253 | 1384 |
| DTBench | 50.9% | 68.3% |
| LMCA | 2.8% | 25.6% |
| Epoch Capabilities Index | 116.02 | 136.79 |
| ARC-AGI-2 | — | 0.4% |
| SimpleBench | — | 27% |
| ARC-AGI-1 | — | 5.5% |
| EnigmaEval | — | 2.2% |
| ForecastBench | — | 61.5 |
Math GPT-4.1 leads
Gemma 3 4B: 15.8 (#286), GPT-4.1: 21.4 (#269)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 38.3% |
| LMArena Math | 1239 | 1369 |
| FrontierMath (Tiers 1-3) | — | 6% |
| Omni-MATH | — | 47.1% |
| MATH Level 5 | — | 83% |
| FrontierMath (Feb 2025 set) | — | 5.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge GPT-4.1 leads
Gemma 3 4B: 11.8 (#294), GPT-4.1: 37.0 (#152)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| GPQA Diamond | 23.2% | 66.9% |
| Vectara Hallucination Rate | 6.4% | 5.6% |
| LMArena Expert | 1223 | 1364 |
| Humanity's Last Exam | — | 5.4% |
| SimpleQA Verified | — | 31.1% |
| MMLU-Pro | — | 81.1% |
| GPQA (HELM) | — | 65.9% |
Multimodal Not comparable
Gemma 3 4B: —, GPT-4.1: 37.5 (#66)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Vision | — | 1210 |
| GeoBench | — | 72% |
Multilingual GPT-4.1 leads
Gemma 3 4B: 42.7 (#189), GPT-4.1: 49.6 (#128)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Non-English | 1273 | 1370 |
| LMArena German | 1281 | 1380 |
| LMArena Russian | 1294 | 1377 |
| LMArena Chinese | — | 1382 |
| LMArena French | — | 1383 |
| LMArena Japanese | — | 1320 |
| LMArena Korean | — | 1339 |
| LMArena Spanish | — | 1375 |
Instruction Following GPT-4.1 leads
Gemma 3 4B: 65.4 (#219), GPT-4.1: 71.5 (#148)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1367 |
| IFEval | — | 83.8% |
Long Context GPT-4.1 leads
Gemma 3 4B: 38.6 (#189), GPT-4.1: 40.0 (#158)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Longer Query | 1274 | 1385 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference GPT-4.1 leads
Gemma 3 4B: 42.2 (#234), GPT-4.1: 57.8 (#120)
| Benchmark | Gemma 3 4B | GPT-4.1 |
|---|---|---|
| LMArena Text | 1291 | 1383 |
| LMArena Creative Writing | 1270 | 1363 |
| EQ-Bench Creative Writing | 1068 | 1420 |
| LMArena Multi-Turn | 1255 | 1398 |
| WildBench | — | 85.4% |
Frequently asked questions
Is Gemma 3 4B better than GPT-4.1?
GPT-4.1 is the stronger model overall, scoring 35.7 to 27.9 on the Noometry Index. Gemma 3 4B costs 70× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or GPT-4.1?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; GPT-4.1 lists at $2 and $8.
Is Gemma 3 4B or GPT-4.1 better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.3 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 131K.
How many benchmarks do Gemma 3 4B and GPT-4.1 share?
22 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and GPT-4.1 has 52.