Description
https://github.com/anomalyco/opencode/issues/new
Description
The OpenCode Zen/Go chat completions API is broken. All POST requests to /zen/v1/chat/completions return HTTP 500 Internal Server Error, regardless of model or API key used.
Steps to Reproduce
- Send POST request to
https://opencode.ai/zen/v1/chat/completions
- Include a valid OpenCode Go subscription API key (
sk-...) in the Authorization header
- Set body to any model (tested:
deepseek-v4-flash-free, minimax-m2.7, gpt-4o-mini)
- Response: HTTP 500 with body
{"type":"error","error":{"type":"error","message":"Internal server error"}}
What Works
GET https://opencode.ai/zen/v1/models — returns 200 with full model list (50 models)
GET https://opencode.ai/zen/go/v1/models — returns 200 with Go model list (20 models)
https://opencode.ai homepage — loads fine
What Doesn't Work
POST https://opencode.ai/zen/v1/chat/completions — returns HTTP 500 (all models, both API keys)
POST https://opencode.ai/zen/go/v1/chat/completions — returns HTTP 000 / connection failure
POST https://opencode.ai/v1/chat/completions — returns HTTP 404
Notes
- Two different OpenCode Go subscription keys were tested, both produce the same 500 error
- Even the free model (
deepseek-v4-flash-free) returns 500
- The model listing endpoints work fine, suggesting only the inference backend is affected
- API keys format: correct (
sk- prefix), freshly generated
Plugins
No response
OpenCode version
No response
Steps to reproduce
No response
Screenshot and/or share link
https://github.com/anomalyco/opencode/issues/new
Operating System
Windows11
Terminal
Windows cmd PowerShell
Description
https://github.com/anomalyco/opencode/issues/new
Description
The OpenCode Zen/Go chat completions API is broken. All POST requests to
/zen/v1/chat/completionsreturn HTTP 500 Internal Server Error, regardless of model or API key used.Steps to Reproduce
https://opencode.ai/zen/v1/chat/completionssk-...) in the Authorization headerdeepseek-v4-flash-free,minimax-m2.7,gpt-4o-mini){"type":"error","error":{"type":"error","message":"Internal server error"}}What Works
GET https://opencode.ai/zen/v1/models— returns 200 with full model list (50 models)GET https://opencode.ai/zen/go/v1/models— returns 200 with Go model list (20 models)https://opencode.aihomepage — loads fineWhat Doesn't Work
POST https://opencode.ai/zen/v1/chat/completions— returns HTTP 500 (all models, both API keys)POST https://opencode.ai/zen/go/v1/chat/completions— returns HTTP 000 / connection failurePOST https://opencode.ai/v1/chat/completions— returns HTTP 404Notes
deepseek-v4-flash-free) returns 500sk-prefix), freshly generatedPlugins
No response
OpenCode version
No response
Steps to reproduce
No response
Screenshot and/or share link
https://github.com/anomalyco/opencode/issues/new
Operating System
Windows11
Terminal
Windows cmd PowerShell