Qwen3-32B is a large language model from Alibaba's Qwen3 series. It features 32.8 billion parameters, a 128k token context window, support for 119 languages, and hybrid thinking modes allowing switching between deep reasoning and fast responses. It demonstrates strong performance in reasoning, instruction-following, and agent capabilities.
发布日期2025年4月29日
参数规模32.8B
上下文长度131K
许可证Apache 2.0
知识截止—
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| Arena Hard | general reasoning creativity | 93.8 | 来源 |
| AIME 2024 | math reasoning | 81.4 | 来源 |
| LiveBench | math reasoning general | 74.9 | 来源 |
| MultiLF | general | 73.0 | 来源 |
| AIME 2025 | math reasoning | 72.9 | 来源 |
| BFCL | general reasoning | 70.3 | 来源 |
| CodeForces | math reasoning | 65.9 | 来源 |
| LiveCodeBench | reasoning general code | 65.7 | 来源 |
| Aider | reasoning code | 50.2 | 来源 |
Pricing
API 价格对比
| 服务商 | 输入价 | 输出价 | 上下文 | 吞吐(tok/s) | 延迟(s) | 函数调用 | 代码执行 | 联网搜索 |
|---|---|---|---|---|---|---|---|---|
| iFlow | $0.00 | $0.00 | 128K | — | — | ✓ | ✗ | ✗ |
| OVHcloud AI Endpoints | $0.09 | $0.25 | 33K | — | — | ✓ | ✗ | ✗ |
| Cortecs | $0.10 | $0.33 | 16K | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $0.10 | $0.30 | 33K | — | — | ✓ | ✗ | ✗ |
| Alibaba (China) | $0.29 | $1.15 | 131K | — | — | ✓ | ✗ | ✗ |
| Helicone | $0.29 | $0.59 | 131K | — | — | ✓ | ✗ | ✗ |
| Alibaba | $0.70 | $2.80 | 131K | — | — | ✓ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。