Bug: Qwen-R1 context window limited to 8K instead of official 16,384 #930

Closed
opened 2026-02-16 10:21:02 -05:00 by yindo · 0 comments
Owner

Originally created by @hsiong on GitHub (Jan 11, 2026).

Self Checks

  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues Dify issues & Dify Official Plugins, including closed ones.
  • I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
  • [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:)
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.11.2

Plugin version

langgenius/tongyi 0.1.18

Cloud or Self Hosted

Cloud

Steps to reproduce

  1. Install Tongyi plugin v0.1.18 from Dify Marketplace.
  2. Configure an LLM using DeepSeek-R1 in the Tongyi plugin.
  3. Select CHAT mode.
  4. Observe the model context length shown in Dify UI or enforced by the plugin.
  5. Compare the configured / displayed context window with the official DeepSeek-R1 specification.

✔️ Error log

No explicit runtime error.
This is a configuration / capability mismatch issue.

The Tongyi plugin incorrectly reports/enforces the context window for Qwen-R1 as 8K tokens, while the official model specification supports 16,384 tokens (16K).

As a result, long prompts or conversations are truncated earlier than expected.

Image

The model details shown in the Tongyi/Dify UI on the Aliyun Bailian console also list DeepSeek-R1

Image
Originally created by @hsiong on GitHub (Jan 11, 2026). ### Self Checks - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing issues [Dify issues](https://github.com/langgenius/dify/issues) & [Dify Official Plugins](https://github.com/langgenius/dify-official-plugins/issues), including closed ones. - [x] I confirm that I am using English to submit this report (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:) - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.11.2 ### Plugin version langgenius/tongyi 0.1.18 ### Cloud or Self Hosted Cloud ### Steps to reproduce 1. Install **Tongyi plugin v0.1.18** from Dify Marketplace. 2. Configure an LLM using **DeepSeek-R1** in the Tongyi plugin. 3. Select **CHAT** mode. 4. Observe the model context length shown in Dify UI or enforced by the plugin. 5. Compare the configured / displayed context window with the official DeepSeek-R1 specification. ### ✔️ Error log No explicit runtime error. This is a configuration / capability mismatch issue. The Tongyi plugin incorrectly reports/enforces the context window for Qwen-R1 as 8K tokens, while the official model specification supports 16,384 tokens (16K). As a result, long prompts or conversations are truncated earlier than expected. <img width="958" height="223" alt="Image" src="https://github.com/user-attachments/assets/342d3a8a-f5a3-4999-96c0-30dae943596f" /> [The model details shown in the Tongyi/Dify UI on the Aliyun Bailian console also list DeepSeek-R1](https://bailian.console.aliyun.com/?spm=5176.29597918.J_SEsSjsNv72yRuRFS2VknO.2.52007b08jZ957l&tab=api#/api/?type=model&url=2868565) <img width="2028" height="727" alt="Image" src="https://github.com/user-attachments/assets/c438775d-cf26-4a2e-bf9d-bda29fb4cf0e" />
yindo added the bug label 2026-02-16 10:21:02 -05:00
yindo closed this issue 2026-02-16 10:21:02 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify-official-plugins#930