Model comparison
Gemma 3 4B vs GPT-3.5-turbo
Gemma 3 4B is the stronger model overall, scoring 28.1 to 23.2 on the Noometry Index.
Last verified . 19 shared benchmarks.
GoogleSummary
- They share 19 benchmarks with published results for both. Gemma 3 4B scores higher in 7 categories and GPT-3.5-turbo in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 3 4B leads 42.0 to 25.3.
- The biggest single-benchmark swing is LMCA: 2.8% for Gemma 3 4B and 9.7% for GPT-3.5-turbo.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.50 / $1.50 for GPT-3.5-turbo.
- Gemma 3 4B accepts more context: 131K tokens versus 16K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 4B | GPT-3.5-turbo | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 28.1 | 23.2 |
| Released | 2025-03-12 | 2023-03-01 |
| Weights | Open | Proprietary |
| Context window | 131K | 16K |
| Max output | 4K | 4K |
| Input $ / M tokens | $0.04 | $0.50 |
| Output $ / M tokens | $0.08 | $1.50 |
| Results tracked | 22 | 44 |
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-3.5-turbo: 23.9 (#331)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| LMArena Coding | 1230 | 1136 |
| WeirdML | — | 3.5% |
| BigCodeBench Instruct | — | 39.1% |
| BigCodeBench Complete | — | 50.6% |
| HumanEval+ | — | 70.7% |
| MBPP+ | — | 69.7% |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.9 (#142), GPT-3.5-turbo: —
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
| METR Time Horizons | — | 21.5% |
Reasoning Too close to call
Gemma 3 4B: 13.2 (#335), GPT-3.5-turbo: 13.8 (#332)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1253 | 1108 |
| DTBench | 50.9% | 48.5% |
| LMCA | 2.8% | 9.7% |
| Epoch Capabilities Index | 116.02 | 118.55 |
| Kagi LLM Benchmark | 25.2% | — |
| Mystery Game Puzzles | — | 3% |
| Adversarial NLI | — | 58.1% |
| BIG-Bench Hard | — | 61.6% |
| CommonsenseQA 2.0 | — | 57% |
| ForecastBench | — | 50.4 |
| WinoGrande | — | 81.6% |
Math Gemma 3 4B leads
Gemma 3 4B: 16.8 (#292), GPT-3.5-turbo: 6.3 (#327)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 2.2% |
| LMArena Math | 1239 | 1142 |
| FrontierMath (Tiers 1-3) | — | 0% |
| MATH Level 5 | — | 15.9% |
| GSM8K | — | 57.8% |
Knowledge Gemma 3 4B leads
Gemma 3 4B: 11.8 (#299), GPT-3.5-turbo: 10.0 (#303)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| GPQA Diamond | 23.2% | 28% |
| LMArena Expert | 1223 | 1070 |
| Vectara Hallucination Rate | 6.4% | — |
| ARC (AI2) Challenge | — | 87.4% |
| BoolQ | — | 87% |
| MMLU | — | 71.4% |
| OpenBookQA | — | 86% |
| TriviaQA | — | 85.8% |
Multilingual Gemma 3 4B leads
Gemma 3 4B: 42.5 (#194), GPT-3.5-turbo: 31.5 (#258)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| LMArena Non-English | 1273 | 1108 |
| LMArena German | 1281 | 1090 |
| LMArena Russian | 1294 | 1123 |
| LMArena Chinese | — | 1075 |
| LMArena French | — | 1118 |
| LMArena Japanese | — | 1043 |
| LMArena Korean | — | 1019 |
| LMArena Spanish | — | 1121 |
Instruction Following Gemma 3 4B leads
Gemma 3 4B: 65.2 (#225), GPT-3.5-turbo: 57.9 (#262)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| LMArena Instruction Following | 1239 | 1119 |
Long Context Gemma 3 4B leads
Gemma 3 4B: 38.7 (#194), GPT-3.5-turbo: 34.0 (#254)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| LMArena Longer Query | 1273 | 1121 |
Writing & Preference Gemma 3 4B leads
Gemma 3 4B: 42.0 (#239), GPT-3.5-turbo: 25.3 (#305)
| Benchmark | Gemma 3 4B | GPT-3.5-turbo |
|---|---|---|
| LMArena Text | 1291 | 1125 |
| LMArena Creative Writing | 1271 | 1092 |
| EQ-Bench Creative Writing | 1068 | 451 |
| LMArena Multi-Turn | 1255 | 1117 |
Frequently asked questions
Is Gemma 3 4B better than GPT-3.5-turbo?
Gemma 3 4B is the stronger model overall, scoring 28.1 to 23.2 on the Noometry Index.
Which is cheaper, Gemma 3 4B or GPT-3.5-turbo?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; GPT-3.5-turbo lists at $0.50 and $1.50.
Is Gemma 3 4B or GPT-3.5-turbo better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 23.9 in the Noometry coding category.
Which has the bigger context window?
Gemma 3 4B does, with 131K tokens against 16K.
How many benchmarks do Gemma 3 4B and GPT-3.5-turbo share?
19 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and GPT-3.5-turbo has 44.