langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-21 03:55:25 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

Chat Assistant qwen3-vl-plus model fails to recognize image content #21503

New Issue
Closed
opened 2026-02-21 20:12:52 -05:00 by yindo · 4 comments
yindo commented 2026-02-21 20:12:52 -05:00
Owner
Copy Link
Copy Source

Originally created by @shifengkuan on GitHub (Jan 7, 2026).

Self Checks

  • I have read the Contributing Guide and Language Policy.
  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report, otherwise it will be closed.
  • 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.11.2

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

When using version 1.11.2, the Chat Assistant, Chatflow, and Agent all utilize Alibaba's qwen3-vl-plus model. While Chatflow and Agent can parse and process images correctly, the Chat Assistant is the only component that fails to recognize image content. What could be the cause?

Image Image Image

✔️ Expected Behavior

What could be the cause?

❌ Actual Behavior

No response

Originally created by @shifengkuan on GitHub (Jan 7, 2026). ### Self Checks - [x] I have read the [Contributing Guide](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542). - [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 [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [x] I confirm that I am using English to submit this report, otherwise it will be closed. - [x] 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：） - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.11.2 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce When using version 1.11.2, the Chat Assistant, Chatflow, and Agent all utilize Alibaba's qwen3-vl-plus model. While Chatflow and Agent can parse and process images correctly, the Chat Assistant is the only component that fails to recognize image content. What could be the cause? <img width="1176" height="312" alt="Image" src="https://github.com/user-attachments/assets/46347d88-0b90-4378-8d79-c8002f62b7b2" /> <img width="939" height="354" alt="Image" src="https://github.com/user-attachments/assets/fcdd695a-0dc6-499a-b4c9-3bc19ca929a5" /> <img width="1183" height="333" alt="Image" src="https://github.com/user-attachments/assets/d025d318-eb77-4cb9-a8ef-420b2d1b1109" /> ### ✔️ Expected Behavior What could be the cause? ### ❌ Actual Behavior _No response_
yindo added the 🐞 bug label 2026-02-21 20:12:52 -05:00
yindo closed this issue 2026-02-21 20:12:52 -05:00
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@dosubot[bot] commented on GitHub (Jan 7, 2026):

The most likely cause is a regression in Dify v1.11.2: image input in the Chat Assistant now requires the model schema to explicitly advertise the VISION feature. Even if the UI shows image upload and the visual toggle is enabled, image input will not work unless the qwen3-vl-plus model's features list includes VISION—there is no manual override for this in 1.11.2. Chatflow and Agent use different config paths and may not enforce this as strictly, which is why they still work.

To resolve this, update your model/provider schema so the features list contains VISION, then restart your provider/plugin and Dify API services. Also, double-check that your INTERNAL_FILES_URL and MULTIMODAL_SEND_FORMAT environment variables are set correctly for Docker deployments, and that file upload is enabled in your app's features. After making changes, start a new conversation for the settings to take effect. If you still have issues, capture backend error logs or API responses when submitting images to help pinpoint the problem.
Reference: discussion and troubleshooting checklist

To reply, just mention @dosu.

How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other

Chat with Dosu Open in Cursor Join Discord Share on X

@dosubot[bot] commented on GitHub (Jan 7, 2026): <!-- Answer --> The most likely cause is a regression in Dify v1.11.2: image input in the Chat Assistant now requires the model schema to explicitly advertise the `VISION` feature. Even if the UI shows image upload and the visual toggle is enabled, image input will not work unless the qwen3-vl-plus model's features list includes `VISION`—there is no manual override for this in 1.11.2. Chatflow and Agent use different config paths and may not enforce this as strictly, which is why they still work. To resolve this, update your model/provider schema so the features list contains `VISION`, then restart your provider/plugin and Dify API services. Also, double-check that your `INTERNAL_FILES_URL` and `MULTIMODAL_SEND_FORMAT` environment variables are set correctly for Docker deployments, and that file upload is enabled in your app's features. After making changes, start a new conversation for the settings to take effect. If you still have issues, capture backend error logs or API responses when submitting images to help pinpoint the problem. [Reference: discussion and troubleshooting checklist](https://github.com/langgenius/dify/discussions/30504#discussioncomment-15411024) <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/bfa38c98-abf1-41ef-b4e8-d0ac9787ce2e?feedback_type=other)</sup> [![Chat with Dosu](https://dosu.dev/dosu-chat-badge.svg)](https://app.dosu.dev/6088a4f4-073a-43f3-b3d4-4f06fb0c05a0/ask?utm_source=github)&nbsp;[![Open in Cursor](https://dosu.dev/dosu-in-cursor.svg)](https://cursor.com/link/prompt?text=The%20most%20likely%20cause%20is%20a%20regression%20in%20Dify%20v1.11.2%3A%20image%20input%20in%20the%20Chat%20Assistant%20now%20requires%20the%20model%20schema%20to%20explicitly%20advertise%20the%20%60VISION%60%20feature.%20Even%20if%20the%20UI%20shows%20image%20upload%20and%20the%20visual%20toggle%20is%20enabled%2C%20image%20input%20will%20not%20work%20unless%20the%20qwen3-vl-plus%20model%27s%20features%20list%20includes%20%60VISION%60%E2%80%94there%20is%20no%20manual%20override%20for%20this%20in%201.11.2.%20Chatflow%20and%20Agent%20use%20different%20config%20paths%20and%20may%20not%20enforce%20this%20as%20strictly%2C%20which%20is%20why%20they%20still%20work.%0A%0ATo%20resolve%20this%2C%20update%20your%20model/provider%20schema%20so%20the%20features%20list%20contains%20%60VISION%60%2C%20then%20restart%20your%20provider/plugin%20and%20Dify%20API%20services.%20Also%2C%20double-check%20that%20your%20%60INTERNAL_FILES_URL%60%20and%20%60MULTIMODAL_SEND_FORMAT%60%20environment%20variables%20are%20set%20correctly%20for%20Docker%20deployments%2C%20and%20that%20file%20upload%20is%20enabled%20in%20your%20app%27s%20features.%20After%20making%20changes%2C%20start%20a%20new%20conversation%20for%20the%20settings%20to%20take%20effect.%20If%20you%20still%20have%20issues%2C%20capture%20backend%20error%20logs%20or%20API%20responses%20when%20submitting%20images%20to%20help%20pinpoint%20the%20problem.%20%20%0A%5BReference%3A%20discussion%20and%20troubleshooting%20checklist%5D%28https%3A//github.com/langgenius/dify/discussions/30504%23discussioncomment-15411024%29)&nbsp;[![Join Discord](https://img.shields.io/badge/join-5865F2?logo=discord&logoColor=white&label=)](https://go.dosu.dev/discord-bot)&nbsp;[![Share on X](https://img.shields.io/badge/X-share-black)](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/30693)
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@shifengkuan commented on GitHub (Jan 7, 2026):

@dosu
How can I update your model/provider settings? I have already upgraded the langgenius/tongyi image to the latest version, and the INTERNAL_FILES_URL and MULTIMODAL_SEND_FORMAT configurations are correct. However, I do not have access to modify the model provider configurations on my end.

Image Image Image
@shifengkuan commented on GitHub (Jan 7, 2026): @dosu How can I update your model/provider settings? I have already upgraded the langgenius/tongyi image to the latest version, and the INTERNAL_FILES_URL and MULTIMODAL_SEND_FORMAT configurations are correct. However, I do not have access to modify the model provider configurations on my end. <img width="219" height="38" alt="Image" src="https://github.com/user-attachments/assets/d4a26cb9-822e-4c03-bdf5-9023e54c4292" /> <img width="338" height="28" alt="Image" src="https://github.com/user-attachments/assets/54a83785-5cfd-4b7e-aa2a-b5c5dcc6c5fd" /> <img width="292" height="137" alt="Image" src="https://github.com/user-attachments/assets/d3984946-abcb-4733-89e0-b9b72447cce2" />
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@dosubot[bot] commented on GitHub (Jan 7, 2026):

When the model is set to “managed by model provider,” you cannot update the model’s features (like adding VISION) from the Dify UI or your own config. Only the provider maintainer or admin can change the model schema to advertise VISION—this is required for image input to work in Chat Assistant as of v1.11.2. Upgrading the langgenius/tongyi image and setting environment variables like INTERNAL_FILES_URL and MULTIMODAL_SEND_FORMAT are necessary, but they do not affect the model’s advertised features.

If you need image input in Chat Assistant, you’ll need to ask whoever manages the provider/plugin to update the model’s features list to include VISION and then restart the provider and Dify API services. There’s currently no user-side override or workaround for this restriction in v1.11.2. If you have access to a different model or provider that already exposes VISION, you can switch to that as an alternative.

If you escalate, provide the provider maintainer with the model name, your Dify version, and a request to ensure the features list includes VISION for the qwen3-vl-plus model.

To reply, just mention @dosu.

How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other

Chat with Dosu Open in Cursor Join Discord Share on X

@dosubot[bot] commented on GitHub (Jan 7, 2026): <!-- Answer --> When the model is set to “managed by model provider,” you cannot update the model’s features (like adding `VISION`) from the Dify UI or your own config. Only the provider maintainer or admin can change the model schema to advertise `VISION`—this is required for image input to work in Chat Assistant as of v1.11.2. Upgrading the langgenius/tongyi image and setting environment variables like `INTERNAL_FILES_URL` and `MULTIMODAL_SEND_FORMAT` are necessary, but they do not affect the model’s advertised features. If you need image input in Chat Assistant, you’ll need to ask whoever manages the provider/plugin to update the model’s features list to include `VISION` and then restart the provider and Dify API services. There’s currently no user-side override or workaround for this restriction in v1.11.2. If you have access to a different model or provider that already exposes `VISION`, you can switch to that as an alternative. If you escalate, provide the provider maintainer with the model name, your Dify version, and a request to ensure the features list includes `VISION` for the qwen3-vl-plus model. <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/4e01dd35-136c-4539-a5f3-9d0ca1f54e23?feedback_type=other)</sup> [![Chat with Dosu](https://dosu.dev/dosu-chat-badge.svg)](https://app.dosu.dev/6088a4f4-073a-43f3-b3d4-4f06fb0c05a0/ask?utm_source=github)&nbsp;[![Open in Cursor](https://dosu.dev/dosu-in-cursor.svg)](https://cursor.com/link/prompt?text=When%20the%20model%20is%20set%20to%20%E2%80%9Cmanaged%20by%20model%20provider%2C%E2%80%9D%20you%20cannot%20update%20the%20model%E2%80%99s%20features%20%28like%20adding%20%60VISION%60%29%20from%20the%20Dify%20UI%20or%20your%20own%20config.%20Only%20the%20provider%20maintainer%20or%20admin%20can%20change%20the%20model%20schema%20to%20advertise%20%60VISION%60%E2%80%94this%20is%20required%20for%20image%20input%20to%20work%20in%20Chat%20Assistant%20as%20of%20v1.11.2.%20Upgrading%20the%20langgenius/tongyi%20image%20and%20setting%20environment%20variables%20like%20%60INTERNAL_FILES_URL%60%20and%20%60MULTIMODAL_SEND_FORMAT%60%20are%20necessary%2C%20but%20they%20do%20not%20affect%20the%20model%E2%80%99s%20advertised%20features.%0A%0AIf%20you%20need%20image%20input%20in%20Chat%20Assistant%2C%20you%E2%80%99ll%20need%20to%20ask%20whoever%20manages%20the%20provider/plugin%20to%20update%20the%20model%E2%80%99s%20features%20list%20to%20include%20%60VISION%60%20and%20then%20restart%20the%20provider%20and%20Dify%20API%20services.%20There%E2%80%99s%20currently%20no%20user-side%20override%20or%20workaround%20for%20this%20restriction%20in%20v1.11.2.%20If%20you%20have%20access%20to%20a%20different%20model%20or%20provider%20that%20already%20exposes%20%60VISION%60%2C%20you%20can%20switch%20to%20that%20as%20an%20alternative.%0A%0AIf%20you%20escalate%2C%20provide%20the%20provider%20maintainer%20with%20the%20model%20name%2C%20your%20Dify%20version%2C%20and%20a%20request%20to%20ensure%20the%20features%20list%20includes%20%60VISION%60%20for%20the%20qwen3-vl-plus%20model.)&nbsp;[![Join Discord](https://img.shields.io/badge/join-5865F2?logo=discord&logoColor=white&label=)](https://go.dosu.dev/discord-bot)&nbsp;[![Share on X](https://img.shields.io/badge/X-share-black)](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/30693)
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@crazywoola commented on GitHub (Jan 7, 2026):

If you’re experiencing a plugin-related issue, please report it using this template: https://github.com/langgenius/dify-official-plugins/issues/new?template=bug_report.yml.

@crazywoola commented on GitHub (Jan 7, 2026): If you’re experiencing a plugin-related issue, please report it using this template: [https://github.com/langgenius/dify-official-plugins/issues/new?template=bug_report.yml](https://github.com/langgenius/dify-official-plugins/issues/new?template=bug_report.yml).
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label 🐞 bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#21503
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?