Repository navigation
[BUG] Azure gpt-5.2-codex (vision-capable) is unable to read screenshots #10150
Description
Description
Bug
When using an Azure-hosted gpt-5.2-codex model in OpenCode TUI, image inputs do not appear to reach the model.
The UI shows an attached/pasted image (e.g. img / clipboard), but the model only receives a textual placeholder like [Image 1] and responds as if no image was provided.
This looks like OpenCode request-shaping / routing dropping image parts for this model/provider combination.
Environment
- OpenCode version: 1.1.33
- OS: macOS (26.2)
- Interface: OpenCode TUI
- Provider: Azure
- Model: gpt-5.2-codex (vision-capable)
Expected
The model describes what is in the screenshot.
Actual
The model behaves as if no image was provided (only a placeholder like [Image 1]).
OpenCode version
1.1.33
Steps to reproduce
- Start OpenCode TUI.
- Select the Azure
gpt-5.2-codexmodel. - Paste or attach a screenshot in the input (confirm the UI indicates an image is attached).
- Ask the model to describe the screenshot.
Screenshot and/or share link
Operating System
macOS 26.2
Terminal
Warp
Activity
github-actions commentedon Jan 23, 2026
This issue might be a duplicate of existing issues. Please check:
Feel free to ignore if none of these address your specific case.
github-actions commentedon Mar 28, 2026
To stay organized issues are automatically closed after 90 days of no activity. If the issue is still relevant please open a new one.