Model comparison
Gemma 3 4B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 28.1 on the Noometry Index. Gemma 3 4B costs 12× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Last verified . 16 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 16 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and Qwen3.8 27B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.8 27B leads 41.6 to 11.8.
- The biggest single-benchmark swing is LMCA: 2.8% for Gemma 3 4B and 41.4% for Qwen3.8 27B.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.04 / $2.30 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 131K.
Side by side
| Gemma 3 4B | Qwen3.8 27B | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 28.1 | 46.0 |
| Released | 2025-03-12 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 4K | 33K |
| Input $ / M tokens | $0.04 | $0.04 |
| Output $ / M tokens | $0.08 | $2.30 |
| Results tracked | 22 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Gemma 3 4B: 35.9 (#215), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1230 | 1482 |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
Agentic & Tool Use Qwen3.8 27B leads
Gemma 3 4B: 20.9 (#142), Qwen3.8 27B: 32.9 (#57)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| Berkeley Function Calling Leaderboard | 19.6% | — |
Reasoning Qwen3.8 27B leads
Gemma 3 4B: 13.2 (#335), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1253 | 1460 |
| DTBench | 50.9% | 88% |
| LMCA | 2.8% | 41.4% |
| Epoch Capabilities Index | 116.02 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| Kagi LLM Benchmark | 25.2% | — |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| Chess Puzzles | 0% | — |
| Surface Evolver Bench | — | 45% |
Math Qwen3.8 27B leads
Gemma 3 4B: 16.8 (#292), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1239 | 1456 |
| OTIS Mock AIME 2024-2025 | 7.5% | — |
| ProofBench | — | 16% |
Knowledge Qwen3.8 27B leads
Gemma 3 4B: 11.8 (#299), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1223 | 1482 |
| GPQA Diamond | 23.2% | — |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
Gemma 3 4B: 42.5 (#194), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1273 | 1430 |
| LMArena German | 1281 | 1438 |
| LMArena Russian | 1294 | 1415 |
| LMArena Chinese | — | 1504 |
| LMArena French | — | 1465 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Qwen3.8 27B leads
Gemma 3 4B: 65.2 (#225), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1239 | 1439 |
Long Context Qwen3.8 27B leads
Gemma 3 4B: 38.7 (#194), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1273 | 1450 |
Writing & Preference Qwen3.8 27B leads
Gemma 3 4B: 42.0 (#239), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Gemma 3 4B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1291 | 1441 |
| LMArena Creative Writing | 1271 | 1384 |
| EQ-Bench Creative Writing | 1068 | 1671 |
| LMArena Multi-Turn | 1255 | 1441 |
Frequently asked questions
Is Gemma 3 4B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 28.1 on the Noometry Index. Gemma 3 4B costs 12× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or Qwen3.8 27B?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Qwen3.8 27B lists at $0.04 and $2.30.
Is Gemma 3 4B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 35.9 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 131K.
How many benchmarks do Gemma 3 4B and Qwen3.8 27B share?
16 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Qwen3.8 27B has 31.