Model

Qwen3-235B-A22B-Instruct-2507

Alibaba Cloud / Qwen Team
开源 Apache 2.0

Qwen3-235B-A22B-Instruct-2507 is the updated instruct version of Qwen3-235B-A22B featuring significant improvements in general capabilities including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage. It provides substantial gains in long-tail knowledge coverage across multiple languages and markedly better alignment with user preferences in subjective and open-ended tasks.

发布日期2025年7月22日
参数规模235B
上下文长度262K
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
ZebraLogic reasoning 95.0 来源
MMLU-Redux language reasoning math general 93.1 来源
IFEval general 88.7 来源
MultiPL-E general language 87.9 来源
Creative Writing v3 creativity writing 87.5 来源
WritingBench writing creativity communication 85.2 来源
CSimpleQA general language 84.3 来源
MMLU-Pro language reasoning math general 83.0 来源
INCLUDE general 79.5 来源
MMLU-ProX language reasoning math general 79.4 来源
Arena-Hard v2 general reasoning creativity 79.2 来源
Multi-IF reasoning communication language 77.5 来源
GPQA reasoning general 77.5 来源
LiveBench 20241125 math reasoning general 75.4 来源
Tau2 retail communication reasoning 71.3 来源
BFCL-v3 general reasoning 70.9 来源
AIME 2025 math reasoning 70.3 来源
SuperGPQA reasoning general math legal healthcare finance chemistry economics physics 62.6 来源
Aider-Polyglot general code 57.3 来源
HMMT25 math 55.4 来源
SimpleQA general reasoning 54.3 来源
LiveCodeBench v6 reasoning general 51.8 来源
PolyMATH math reasoning spatial_reasoning multimodal vision 50.2 来源
Tau2 airline reasoning communication 44.0 来源
ARC-AGI reasoning vision spatial_reasoning 41.8 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
Cortecs $0.06 $0.41 131K
LLM Gateway $0.09 $0.58 262K
Venice AI $0.15 $0.75 128K
302.AI $0.29 $1.14 128K
Scaleway $0.75 $2.25 260K

价格单位:美元/百万 token,数据来自社区整理,仅供参考。