Model comparison
Gemma 3 4B vs Qwen3.5-Flash
Qwen3.5-Flash is the stronger model overall, scoring 42.5 to 28.1 on the Noometry Index. Gemma 3 4B costs 3.5× less per token, which makes it the better buy when Qwen3.5-Flash's lead doesn't matter for your workload.
Last verified . 19 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Gemma 3 4B scores higher in 1 category and Qwen3.5-Flash in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.5-Flash leads 43.2 to 11.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 84.4% for Qwen3.5-Flash.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.10 / $0.40 for Qwen3.5-Flash.
- Qwen3.5-Flash 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 | Qwen3.5-Flash | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 28.1 | 42.5 |
| Released | 2025-03-12 | 2026-02-23 |
| Weights | Open | Proprietary |
| Context window | 131K | 1M |
| Max output | 4K | 66K |
| Input $ / M tokens | $0.04 | $0.10 |
| Output $ / M tokens | $0.08 | $0.40 |
| Results tracked | 22 | 32 |
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), Qwen3.5-Flash: 34.2 (#242)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| LMArena Coding | 1230 | 1412 |
| LMArena WebDev | — | 1244 |
| ALE-Bench | — | 221.8 |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.9 (#142), Qwen3.5-Flash: —
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
| Vending-Bench 2 | — | 462.69 |
Reasoning Qwen3.5-Flash leads
Gemma 3 4B: 13.2 (#335), Qwen3.5-Flash: 33.7 (#72)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| Chess Puzzles | 0% | 21% |
| LMArena Hard Prompts | 1253 | 1403 |
| DTBench | 50.9% | 82.9% |
| LMCA | 2.8% | 29.1% |
| Epoch Capabilities Index | 116.02 | 143.98 |
| Kagi LLM Benchmark | 25.2% | — |
| Mystery Game Puzzles | — | 20% |
Math Qwen3.5-Flash leads
Gemma 3 4B: 16.8 (#292), Qwen3.5-Flash: 37.4 (#158)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 84.4% |
| LMArena Math | 1239 | 1407 |
| FrontierMath (Tiers 1-3) | — | 18.2% |
| FrontierMath (Feb 2025 set) | — | 6.2% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Qwen3.5-Flash leads
Gemma 3 4B: 11.8 (#299), Qwen3.5-Flash: 43.2 (#93)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| GPQA Diamond | 23.2% | 82.3% |
| Vectara Hallucination Rate | 6.4% | 10.5% |
| LMArena Expert | 1223 | 1407 |
| SimpleQA Verified | — | 20.3% |
Multilingual Qwen3.5-Flash leads
Gemma 3 4B: 42.5 (#194), Qwen3.5-Flash: 50.5 (#121)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| LMArena Non-English | 1273 | 1385 |
| LMArena German | 1281 | 1390 |
| LMArena Russian | 1294 | 1379 |
| LMArena Chinese | — | 1446 |
| LMArena French | — | 1412 |
| LMArena Japanese | — | 1368 |
| LMArena Korean | — | 1344 |
| LMArena Spanish | — | 1400 |
Instruction Following Qwen3.5-Flash leads
Gemma 3 4B: 65.2 (#225), Qwen3.5-Flash: 72.6 (#139)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| LMArena Instruction Following | 1239 | 1374 |
Long Context Qwen3.5-Flash leads
Gemma 3 4B: 38.7 (#194), Qwen3.5-Flash: 42.4 (#124)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| LMArena Longer Query | 1273 | 1392 |
Writing & Preference Qwen3.5-Flash leads
Gemma 3 4B: 42.0 (#239), Qwen3.5-Flash: 57.9 (#122)
| Benchmark | Gemma 3 4B | Qwen3.5-Flash |
|---|---|---|
| LMArena Text | 1291 | 1397 |
| LMArena Creative Writing | 1271 | 1343 |
| LMArena Multi-Turn | 1255 | 1393 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Qwen3.5-Flash?
Qwen3.5-Flash is the stronger model overall, scoring 42.5 to 28.1 on the Noometry Index. Gemma 3 4B costs 3.5× less per token, which makes it the better buy when Qwen3.5-Flash's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or Qwen3.5-Flash?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Qwen3.5-Flash lists at $0.10 and $0.40.
Is Gemma 3 4B or Qwen3.5-Flash better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5-Flash does, with 1M tokens against 131K.
How many benchmarks do Gemma 3 4B and Qwen3.5-Flash share?
19 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Qwen3.5-Flash has 32.