Model comparison
Gemma 4 31B IT vs Qwen3.5 397B-A17B
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 43.5 on the Noometry Index. Gemma 4 31B IT costs 8.9× less per token, which makes it the better buy when Qwen3.5 397B-A17B's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 26 benchmarks with published results for both. Gemma 4 31B IT scores higher in 5 categories and Qwen3.5 397B-A17B in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.5 397B-A17B leads 53.3 to 37.9.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 73.3% for Gemma 4 31B IT and 88.9% for Qwen3.5 397B-A17B.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $0.60 / $3.60 for Qwen3.5 397B-A17B.
Side by side
| Gemma 4 31B IT | Qwen3.5 397B-A17B | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 43.5 | 46.0 |
| Released | 2026-04-02 | 2026-02-01 |
| Weights | Open | Open |
| Context window | 262K | 262K |
| Max output | 33K | 66K |
| Input $ / M tokens | $0.09 | $0.60 |
| Output $ / M tokens | $0.34 | $3.60 |
| Results tracked | 35 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 4 31B IT: 42.3 (#108), Qwen3.5 397B-A17B: 42.0 (#114)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena WebDev | 1366 | 1400 |
| LMArena Coding | 1459 | 1465 |
| SciCode | 43.4% | — |
| WeirdML | 52.3% | — |
| ALE-Bench | 925.5 | — |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, Qwen3.5 397B-A17B: 33.3 (#53)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| APEX-Agents | — | 24.9% |
| τ²-bench Airline | — | 81.5% |
| τ²-bench Banking | — | 9.8% |
| τ²-bench Retail | — | 84.4% |
| τ²-bench Telecom | — | 97.8% |
Reasoning Qwen3.5 397B-A17B leads
Gemma 4 31B IT: 27.2 (#122), Qwen3.5 397B-A17B: 34.5 (#70)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 73.7% |
| NYT Connections (extended) | 70.6% | 58.9% |
| Chess Puzzles | 5% | 13% |
| Thematic Generalization | 53% | 65.1% |
| LMArena Hard Prompts | 1448 | 1448 |
| DTBench | 82.7% | 87.5% |
| LMCA | 39.3% | 37.9% |
| Epoch Capabilities Index | 142.74 | 146.65 |
| CritPt | 1.4% | — |
| Mystery Game Puzzles | — | 18% |
| Surface Evolver Bench | 30.6% | — |
Math Qwen3.5 397B-A17B leads
Gemma 4 31B IT: 43.2 (#81), Qwen3.5 397B-A17B: 46.1 (#73)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 88.9% |
| LMArena Math | 1465 | 1454 |
| FrontierMath (Tiers 1-3) | — | 31.2% |
Knowledge Qwen3.5 397B-A17B leads
Gemma 4 31B IT: 37.9 (#151), Qwen3.5 397B-A17B: 53.3 (#58)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| GPQA Diamond | 75.8% | 86.4% |
| LMArena Expert | 1465 | 1462 |
| SimpleQA Verified | 10.4% | — |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Too close to call
Gemma 4 31B IT: 41.6 (#34), Qwen3.5 397B-A17B: 40.7 (#44)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Vision | 1277 | 1263 |
| LMArena Document | 1425 | — |
Multilingual Too close to call
Gemma 4 31B IT: 53.8 (#57), Qwen3.5 397B-A17B: 53.7 (#59)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Non-English | 1431 | 1430 |
| LMArena Chinese | 1476 | 1500 |
| LMArena French | 1435 | 1461 |
| LMArena Russian | 1460 | 1429 |
| LMArena Spanish | 1444 | 1441 |
| LMArena German | — | 1447 |
| LMArena Japanese | — | 1426 |
| LMArena Korean | — | 1384 |
Instruction Following Too close to call
Gemma 4 31B IT: 75.5 (#61), Qwen3.5 397B-A17B: 75.0 (#77)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Instruction Following | 1433 | 1424 |
Long Context Too close to call
Gemma 4 31B IT: 44.2 (#71), Qwen3.5 397B-A17B: 44.1 (#74)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Longer Query | 1446 | 1442 |
Writing & Preference Qwen3.5 397B-A17B leads
Gemma 4 31B IT: 60.5 (#96), Qwen3.5 397B-A17B: 62.3 (#79)
| Benchmark | Gemma 4 31B IT | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Text | 1443 | 1438 |
| LMArena Creative Writing | 1415 | 1401 |
| EQ-Bench Creative Writing | 1368 | 1478 |
| LMArena Multi-Turn | 1452 | 1446 |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than Qwen3.5 397B-A17B?
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 43.5 on the Noometry Index. Gemma 4 31B IT costs 8.9× less per token, which makes it the better buy when Qwen3.5 397B-A17B's lead doesn't matter for your workload.
Which is cheaper, Gemma 4 31B IT or Qwen3.5 397B-A17B?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; Qwen3.5 397B-A17B lists at $0.60 and $3.60.
Is Gemma 4 31B IT or Qwen3.5 397B-A17B better for coding?
They score almost the same on coding (42.3 vs 42.0); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 262K tokens.
How many benchmarks do Gemma 4 31B IT and Qwen3.5 397B-A17B share?
26 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Qwen3.5 397B-A17B has 36.