Model comparison
Gemma 3 4B vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 28.1 on the Noometry Index. Gemma 3 4B costs 60× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and Qwen3.8 Max in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 16.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 100% for Qwen3.8 Max.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 4B | Qwen3.8 Max | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 28.1 | 56.8 |
| Released | 2025-03-12 | 2026-08-02 |
| Weights | Open | Proprietary |
| Context window | 131K | 1M |
| Max output | 4K | 131K |
| Input $ / M tokens | $0.04 | $2 |
| Output $ / M tokens | $0.08 | $6 |
| Results tracked | 22 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
Gemma 3 4B: 35.9 (#215), Qwen3.8 Max: 53.5 (#29)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Coding | 1230 | 1502 |
| DeepSWE | — | 57.5% |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| SciCode | — | 53.2% |
Agentic & Tool Use Qwen3.8 Max leads
Gemma 3 4B: 20.9 (#142), Qwen3.8 Max: 45.4 (#14)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | — | 63.3% |
| Berkeley Function Calling Leaderboard | 19.6% | — |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |
Reasoning Qwen3.8 Max leads
Gemma 3 4B: 13.2 (#335), Qwen3.8 Max: 54.4 (#26)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| Chess Puzzles | 0% | 40% |
| LMArena Hard Prompts | 1253 | 1496 |
| DTBench | 50.9% | 92% |
| LMCA | 2.8% | 46.2% |
| Epoch Capabilities Index | 116.02 | 156.41 |
| Kagi LLM Benchmark | 25.2% | — |
| NYT Connections (extended) | — | 88.3% |
| CritPt | — | 20% |
| Mystery Game Puzzles | — | 38% |
Math Qwen3.8 Max leads
Gemma 3 4B: 16.8 (#292), Qwen3.8 Max: 73.2 (#20)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 100% |
| LMArena Math | 1239 | 1499 |
| FrontierMath (Tiers 1-3) | — | 74.7% |
| FrontierMath Tier 4 | — | 46.3% |
| ProofBench | — | 58% |
Knowledge Qwen3.8 Max leads
Gemma 3 4B: 11.8 (#299), Qwen3.8 Max: 61.7 (#27)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 23.2% | 92.7% |
| LMArena Expert | 1223 | 1507 |
| SimpleQA Verified | — | 47.3% |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, Qwen3.8 Max: 37.2 (#75)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | — | 1314 |
| Furniture Assembly | — | 20% |
Multilingual Qwen3.8 Max leads
Gemma 3 4B: 42.5 (#194), Qwen3.8 Max: 56.7 (#18)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1273 | 1472 |
| LMArena German | 1281 | 1483 |
| LMArena Russian | 1294 | 1481 |
| LMArena Chinese | — | 1538 |
| LMArena French | — | 1503 |
| LMArena Japanese | — | 1467 |
| LMArena Korean | — | 1461 |
| LMArena Spanish | — | 1492 |
Instruction Following Qwen3.8 Max leads
Gemma 3 4B: 65.2 (#225), Qwen3.8 Max: 77.6 (#17)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1239 | 1479 |
Long Context Qwen3.8 Max leads
Gemma 3 4B: 38.7 (#194), Qwen3.8 Max: 45.6 (#31)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1273 | 1489 |
Writing & Preference Qwen3.8 Max leads
Gemma 3 4B: 42.0 (#239), Qwen3.8 Max: 67.1 (#30)
| Benchmark | Gemma 3 4B | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1291 | 1483 |
| LMArena Creative Writing | 1271 | 1479 |
| LMArena Multi-Turn | 1255 | 1489 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 28.1 on the Noometry Index. Gemma 3 4B costs 60× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or Qwen3.8 Max?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is Gemma 3 4B or Qwen3.8 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 35.9 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 Max does, with 1M tokens against 131K.
How many benchmarks do Gemma 3 4B and Qwen3.8 Max share?
18 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Qwen3.8 Max has 39.