# 推理模型400错误

**URL:** <https://meta.discourse.org/t/reasoning-models-400-error/349334>\
**Category:** Support\
**Tags:** ai\
**Created:** [2025年一月28日 08:21 UTC](https://meta.discourse.org/t/reasoning-models-400-error/349334 "2025-01-28T08:21:28Z")\
**Posts on this page:** 1\
**Showing post:** 1

<div class="post-metadata">

**Author:** ![MachineScholar](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/machinescholar/32/424719_2.png) [@MachineScholar](https://meta.discourse.org/u/MachineScholar)\
**Post date:** [2025年一月28日 08:21 UTC](https://meta.discourse.org/t/reasoning-models-400-error/349334/1 "2025-01-28T08:21:28Z")

</div>

DeepSeek-R1 在尝试通过 PM 与其交互时产生 400 错误，导致作业异常。

[details=“Backtrace”]

```plaintext
Message (3 copies reported)

DiscourseAi::Completions::Endpoints::OpenAi: status: 400 - body: {"error":{"message":"deepseek-reason
```

---

_[View the full topic](https://meta.discourse.org/t/reasoning-models-400-error/349334)._