Gemma 4 31B vs Qwen3 Next 80B A3B Thinking
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | Gemma 4 31B | Qwen3 Next 80B A3B Thinking |
|---|---|---|
| Provider |  Google DeepMind |  Alibaba Qwen |
| score moyen | 61.6 | 61.6 |
| Prix d'entrée | $0.13 | $0.10 |
| Prix de sortie | $0.38 | $0.78 |
| Fenêtre de contexte | 262K tokens (~131 books) | 131K tokens (~66 books) |
| Sorti le | 2026-04-02 | 2025-09-11 |
| Code source ouvert | Open Source | Open Source |
Scores de benchmark
8 benchmarks · Gemma 4 31B: 6, Qwen3 Next 80B A3B Thinking: 2
| Benchmark | Catégorie | Gemma 4 31B | Qwen3 Next 80B A3B Thinking |
|---|---|---|---|
| LiveBench — Agentic Coding | coding | 40.0 | 8.3 |
| LiveBench — Coding | coding | 60.3 | 60.7 |
| LiveBench — Data Analysis | reasoning | 58.8 | 53.6 |
| LiveBench — If | language | 67.6 | 41.5 |
| LiveBench — Language | language | 71.3 | 56.3 |
| LiveBench — Mathematics | math | 73.9 | 74.3 |
| LiveBench — Overall | knowledge | 61.6 | 50.4 |
| LiveBench — Reasoning | reasoning | 59.4 | 58.2 |