Model comparison
Gemma 3 4B vs Qwen1.5-72B
Qwen1.5-72B is the stronger model overall, scoring 30.8 to 28.1 on the Noometry Index.
Last verified . 13 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. Gemma 3 4B scores higher in 6 categories and Qwen1.5-72B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen1.5-72B leads 33.2 to 16.8.
- The biggest single-benchmark swing is GPQA Diamond: 23.2% for Gemma 3 4B and 28.8% for Qwen1.5-72B.
Side by side
| Gemma 3 4B | Qwen1.5-72B | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 28.1 | 30.8 |
| Released | 2025-03-12 | 2024-02-04 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.04 | — |
| Output $ / M tokens | $0.08 | — |
| Results tracked | 22 | 22 |
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), Qwen1.5-72B: 31.9 (#277)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Coding | 1230 | 1165 |
| BigCodeBench Instruct | — | 33.2% |
| BigCodeBench Complete | — | 40.3% |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 61.6% |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.9 (#142), Qwen1.5-72B: —
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
Reasoning Qwen1.5-72B leads
Gemma 3 4B: 13.2 (#335), Qwen1.5-72B: 22.2 (#203)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Hard Prompts | 1253 | 1148 |
| Kagi LLM Benchmark | 25.2% | — |
| Chess Puzzles | 0% | — |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| Epoch Capabilities Index | 116.02 | — |
Math Qwen1.5-72B leads
Gemma 3 4B: 16.8 (#292), Qwen1.5-72B: 33.2 (#205)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Math | 1239 | 1164 |
| OTIS Mock AIME 2024-2025 | 7.5% | — |
Knowledge Too close to call
Gemma 3 4B: 11.8 (#299), Qwen1.5-72B: 11.5 (#300)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| GPQA Diamond | 23.2% | 28.8% |
| LMArena Expert | 1223 | 1136 |
| Vectara Hallucination Rate | 6.4% | — |
Multilingual Gemma 3 4B leads
Gemma 3 4B: 42.5 (#194), Qwen1.5-72B: 33.2 (#253)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Non-English | 1273 | 1135 |
| LMArena German | 1281 | 1084 |
| LMArena Russian | 1294 | 1104 |
| LMArena Chinese | — | 1186 |
| LMArena French | — | 1159 |
| LMArena Japanese | — | 1061 |
| LMArena Korean | — | 1050 |
| LMArena Spanish | — | 1110 |
Instruction Following Gemma 3 4B leads
Gemma 3 4B: 65.2 (#225), Qwen1.5-72B: 59.3 (#256)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Instruction Following | 1239 | 1141 |
Long Context Gemma 3 4B leads
Gemma 3 4B: 38.7 (#194), Qwen1.5-72B: 35.1 (#243)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Longer Query | 1273 | 1157 |
Writing & Preference Gemma 3 4B leads
Gemma 3 4B: 42.0 (#239), Qwen1.5-72B: 37.3 (#258)
| Benchmark | Gemma 3 4B | Qwen1.5-72B |
|---|---|---|
| LMArena Text | 1291 | 1166 |
| LMArena Creative Writing | 1271 | 1137 |
| LMArena Multi-Turn | 1255 | 1160 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Qwen1.5-72B?
Qwen1.5-72B is the stronger model overall, scoring 30.8 to 28.1 on the Noometry Index.
Is Gemma 3 4B or Qwen1.5-72B better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 31.9 in the Noometry coding category.
How many benchmarks do Gemma 3 4B and Qwen1.5-72B share?
13 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Qwen1.5-72B has 22.