github-starred/open-webui
2
0
Fork 0
mirror of https://github.com/open-webui/open-webui.git synced 2026-07-27 17:04:57 -05:00
Code Issues 2.4k Packages Projects Releases 129 Wiki Activity

[GH-ISSUE #23244] issue: Image generation fails (empty response) when using Gemini as chat model, but works with GPT models #74531

New Issue
Closed
opened 2026-05-13 07:18:08 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-13 07:18:08 -05:00
Owner
Copy Link
Copy Source

Originally created by @Tetsuro-Copa on GitHub (Mar 31, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/23244

Check Existing Issues

  • I have searched for any existing and/or related issues.
  • I have searched for any existing and/or related discussions.
  • I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
  • I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v 0.8.12

Ollama Version (if applicable)

No response

Operating System

Windows 11

Browser (if applicable)

Brave (Chromium)

Confirmation

  • I have read and followed all instructions in README.md.
  • I am using the latest version of both Open WebUI and Ollama.
  • I have included the browser console logs.
  • I have included the Docker container logs.
  • I have provided every relevant configuration, setting, and environment variable used in my setup.
  • I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
  • I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
  • Start with the initial platform/version/OS and dependencies used,
  • Specify exact install/launch/configure commands,
  • List URLs visited, user input (incl. example values/emails/passwords if needed),
  • Describe all options and toggles enabled or changed,
  • Include any files or environmental changes,
  • Identify the expected and actual result at each stage,
  • Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

Expected Behavior

When requesting image generation (e.g. "Generate an image of a cat dancing on a cake"), the system should:

  • Trigger the image generation tool (generate_image)
  • Return a valid response
  • Populate content, output, and files with the generated image

Actual Behavior

Actual Behavior

When using Gemini as the chat model:

  • Assistant message is saved with:
    • content: ""
    • output: []
    • done: true
  • No error is shown in logs
  • No image is generated or displayed
  • Follow-up suggestions still appear

This is a silent failure.

Steps to Reproduce

Description

Image generation fails silently when using a Gemini model (e.g. gemini-2.5-pro) as the chat model in OpenWebUI.

The same image generation setup works correctly when switching the chat model to a GPT-based model (e.g. gpt-5).

Comparison

Working Configuration

  • Chat Model: gpt-5
  • Image Model: gpt-image-1

Result:

  • Tool call (generate_image) is executed
  • Image is generated and stored in files
  • Response is correctly displayed

Failing Configuration

  • Chat Model: gemini-2.5-pro
  • Image Model: gpt-image-1

Result:

  • No tool execution output is recorded
  • Assistant response is empty
  • files is empty

Additional Observations

  • Basic text chat works correctly with Gemini
  • Failures occur only when:
    • Image generation is requested
    • Web search / tool-based features are used
  • Disabling Tools prevents silent failure in some cases, but image generation still fails

Environment

  • OpenWebUI version: v0.8.12
  • Deployment: Docker (Windows host)
  • Reverse proxy: Caddy (LAN only, HTTPS)
  • Browser: Chrome / Brave (both tested)

Hypothesis

The issue appears to be related to how OpenWebUI handles tool/function calls when using Gemini models.

Possibly:

  • Gemini responses are not parsed correctly into tool calls
  • Or the tool execution output is not properly mapped into content / output

Request

  • Is Gemini officially supported for tool-based image generation workflows?
  • Are there known compatibility issues with Gemini and generate_image?
  • Is there a recommended configuration for combining Gemini (chat) with OpenAI (image)?

Any guidance or confirmation would be appreciated.

Logs & Screenshots

Logs

No errors are reported in logs during failure.

Relevant behavior:

  • /api/chat/completions returns 200
  • /api/chat/completed returns 200
  • Final stored message is empty

Additional Information

No response

Originally created by @Tetsuro-Copa on GitHub (Mar 31, 2026). Original GitHub issue: https://github.com/open-webui/open-webui/issues/23244 ### Check Existing Issues - [x] I have searched for any existing and/or related issues. - [x] I have searched for any existing and/or related discussions. - [x] I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!). - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version v 0.8.12 ### Ollama Version (if applicable) _No response_ ### Operating System Windows 11 ### Browser (if applicable) Brave (Chromium) ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior ## Expected Behavior When requesting image generation (e.g. "Generate an image of a cat dancing on a cake"), the system should: - Trigger the image generation tool (`generate_image`) - Return a valid response - Populate `content`, `output`, and `files` with the generated image ### Actual Behavior ## Actual Behavior When using Gemini as the chat model: - Assistant message is saved with: - `content: ""` - `output: []` - `done: true` - No error is shown in logs - No image is generated or displayed - Follow-up suggestions still appear This is a silent failure. ### Steps to Reproduce ## Description Image generation fails silently when using a Gemini model (e.g. `gemini-2.5-pro`) as the chat model in OpenWebUI. The same image generation setup works correctly when switching the chat model to a GPT-based model (e.g. `gpt-5`). ## Comparison ### Working Configuration - Chat Model: `gpt-5` - Image Model: `gpt-image-1` Result: - Tool call (`generate_image`) is executed - Image is generated and stored in `files` - Response is correctly displayed ### Failing Configuration - Chat Model: `gemini-2.5-pro` - Image Model: `gpt-image-1` Result: - No tool execution output is recorded - Assistant response is empty - `files` is empty ## Additional Observations - Basic text chat works correctly with Gemini - Failures occur only when: - Image generation is requested - Web search / tool-based features are used - Disabling Tools prevents silent failure in some cases, but image generation still fails ## Environment - OpenWebUI version: v0.8.12 - Deployment: Docker (Windows host) - Reverse proxy: Caddy (LAN only, HTTPS) - Browser: Chrome / Brave (both tested) ## Hypothesis The issue appears to be related to how OpenWebUI handles tool/function calls when using Gemini models. Possibly: - Gemini responses are not parsed correctly into tool calls - Or the tool execution output is not properly mapped into `content` / `output` ## Request - Is Gemini officially supported for tool-based image generation workflows? - Are there known compatibility issues with Gemini and `generate_image`? - Is there a recommended configuration for combining Gemini (chat) with OpenAI (image)? Any guidance or confirmation would be appreciated. ### Logs & Screenshots ## Logs No errors are reported in logs during failure. Relevant behavior: - `/api/chat/completions` returns 200 - `/api/chat/completed` returns 200 - Final stored message is empty ### Additional Information _No response_
GiteaMirror added the bug label 2026-05-13 07:18:08 -05:00
GiteaMirror commented 2026-05-13 07:18:10 -05:00
Author
Owner
Copy Link
Copy Source

@tjbck commented on GitHub (Apr 2, 2026):

Does gemini-2.5-pro support full openai compatiblity?

<!-- gh-comment-id:4175171351 --> @tjbck commented on GitHub (Apr 2, 2026): Does `gemini-2.5-pro` support full openai compatiblity?
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
bug
confirmed
confirmed issue
core
documentation
enhancement
good first issue
help wanted
non-core
pull-request
Mirrored from GitHub Pull Request
 python
question
testing wanted
No labels bug
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
GiteaMirror ninjasurge (Kaehlen Rog)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#74531