400 Error using ZenMux with Kimi K2.5: "reasoning_content is missing"

[Vwings]

Description

Provider: ZenMux
Model: Kimi K2.5
Error Message:
{"code":"400","type":"invalid_params","message":"thinking is enabled but reasoning_content is missing in assistant tool call message at index 2"}}

Plugins

No response

OpenCode version

1.2.15

Steps to reproduce

1. Configure opencode to use ZenMux provider
2. Select model: Kimi K2.5
3. Send the following prompt to trigger codebase analysis: what's going on in this project

Screenshot and/or share link

No response

Operating System

Windows 11

Terminal

Windows Terminal

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• Kimi K2.5 can't enable thinking (reasoning_content missing #12090: Kimi K2.5 can't enable thinking (reasoning_content missing) — same error and model, now closed
• Why does opencode send reasoning_content in OpenAI-compatible messages, and why does it affect model output? #13130: Why does opencode send reasoning_content in OpenAI-compatible messages — related root cause discussion
• Kimi 2.5 Reasoning Cutoff/Missing when using NVIDIA API #13515: Kimi 2.5 Reasoning Cutoff/Missing when using NVIDIA API — similar reasoning_content handling issue with Kimi K2.5

If your issue is distinct from the above, please add a note explaining the difference.

[OpeOginni]

@Vwings based on this previous solution #12090 (comment), could you try adding this to the model config in your opencode.json?

 "provider": {
    "zenmux": {
      "models": {
        "moonshotai/kimi-k2.5": {
          "options": {
             "thinking": {
                "type": "enabled",
                "budgetTokens": 16000 // this is required for ai sdk even tho it doesnt do anything in actuality
              }
          }
        }
      }
    }
  }

[Vwings]

@Vwings based on this previous solution #12090 (comment), could you try adding this to the model config in your opencode.json?

"provider": {
"zenmux": {
"models": {
"moonshotai/kimi-k2.5": {
"options": {
"thinking": {
"type": "enabled",
"budgetTokens": 16000 // this is required for ai sdk even tho it doesnt do anything in actuality
}
}
}
}
}
}

Tried that, but the issue persists with the exact same error message.