Description
Requests to https://opencode.ai/zen/go/v1/chat/completions specifying "model": "deepseek-v4-flash" fail with an HTTP 400 error originating from the upstream relay validator:
{
"error": {
"type": "invalid_request_error",
"message": "Error from provider (Console Go): Upstream request failed: [invalid_request_error] The supported API model names are deepseek-v4-pro or deepseek-v4-flash, but you passed deepseek-v4-flash."
}
}Key Evidence
Notice the double space in the upstream error message: "...but you passed deepseek-v4-flash.". The OpenCode Go relay prepends a space (" " + model_id) before forwarding the request to Console Go, causing upstream model validation to fail.
Other models on the same endpoint (deepseek-v4-pro, glm-5.2, kimi-k2.5) return HTTP 200 SUCCESS cleanly.
Expected Behavior
deepseek-v4-flash should pass through the OpenCode Go relay without injected whitespace and return HTTP 200 SUCCESS.
Description
Requests to
https://opencode.ai/zen/go/v1/chat/completionsspecifying"model": "deepseek-v4-flash"fail with an HTTP 400 error originating from the upstream relay validator:{ "error": { "type": "invalid_request_error", "message": "Error from provider (Console Go): Upstream request failed: [invalid_request_error] The supported API model names are deepseek-v4-pro or deepseek-v4-flash, but you passed deepseek-v4-flash." } }Key Evidence
Notice the double space in the upstream error message:
"...but you passed deepseek-v4-flash.". The OpenCode Go relay prepends a space (" " + model_id) before forwarding the request to Console Go, causing upstream model validation to fail.Other models on the same endpoint (
deepseek-v4-pro,glm-5.2,kimi-k2.5) return HTTP 200 SUCCESS cleanly.Expected Behavior
deepseek-v4-flashshould pass through the OpenCode Go relay without injected whitespace and return HTTP 200 SUCCESS.