### Description Steps to reproduce: 1. Use OpenCode with any paid provider (GitHub Copilot, Anthropic direct, etc.) 2. Drag/drop an image into the chat 3. Error: "Cannot read 'image.png' (this model does not support image input)" Expected behavior: - Image should be processed correctly (like with free models) Environment: - OS: Windows - Provider used: github-copilot (several models tested: claude-opus-4.6, claude-sonnet-4.5, gpt-5.2) - Free model that works: big-pickle Notes: - Error occurs with ALL paid models tested - Free models like big-pickle DO accept images - Error message says model doesn't support images, but that's false for Claude Opus/Sonnet and GPT-5.2 which are multimodal ### Plugins _No response_ ### OpenCode version _No response_ ### Steps to reproduce _No response_ ### Screenshot and/or share link _No response_ ### Operating System _No response_ ### Terminal _No response_
Description
Steps to reproduce:
Expected behavior:
Environment:
Notes:
Plugins
No response
OpenCode version
No response
Steps to reproduce
No response
Screenshot and/or share link
No response
Operating System
No response
Terminal
No response