编排:Qwen3-max

这不是客观事实。 本页记录本站的梯队、星级、六维和情绪等主观判断，与模型事实页物理分离，仅编排员和管理员可修改。

<syntaxhighlight lang="yaml"> id: qwen3-max tier: T1 stars: 5 composite: 83.6 stats:

 code: 84
 reasoning: 78
 context: 62
 speed: 72
 multimodal: 64
 value: 92

sentiment:

 positive: 45
 mixed: 40
 negative: 15

radar: - axis: 长程任务

 value: 72

- axis: 编程工程

 value: 88

- axis: 抽象推理

 value: 82

- axis: 上下文利用

 value: 70

- axis: 中文能力

 value: 92

- axis: 响应速度

 value: 85

- axis: 稳定性

 value: 65

- axis: 指令遵循

 value: 80

- axis: 易用性

 value: 75

- axis: 性价比

 value: 70

danmaku: - text: Pretty knowledgeable and very fast

 platform: reddit
 main: true

- text: The benchmarks are favouring Qwen3 max

 platform: reddit
 main: true

- text: Qwen is never going to open source Qwen 3.7

 platform: reddit
 main: false

- text: Max was always closed

 platform: reddit
 main: false

- text: No tiene nada de especial q3 max

 platform: reddit
 main: false

- text: Thinking 帖 502 points：scaling works

 platform: hn
 main: true

- text: 正式版发布帖仅 5 分，Thinking 版才引爆

 platform: hn
 main: false

- text: spend more to get more，不是 get more for less

 platform: hn
 main: false

- text: Qwen3时代的最强一击

 platform: zhihu
 main: true

- text: 编程秒了 Opus4，一次成功无需抽卡

 platform: zhihu
 main: false

- text: 中位甚至只提了 9%

 platform: zhihu
 main: false

- text: 写代码建议先别碰这个模型

 platform: zhihu
 main: false

- text: Max推理版用起来费用不低

 platform: zhihu
 main: false

- text: Qoder 里 0.2 倍积分，性价比很高

 platform: v2ex
 main: false

- text: Just Scale it——官方博客标题成社区标签

 platform: x
 main: true

- text: blazing fast，比 ChatGPT 还快

 platform: x
 main: false

</syntaxhighlight>