Agent model validation fails with trailing slash for custom OpenAI-compatible provider models

Description

When using a custom provider model in agent configuration (plan/build/explore), the model validation fails because a trailing slash (/) is appended to the model name during validation.

Error message:

Agent plan's configured model my-custom-model/ is not valid

Note the trailing / at the end - my configuration does NOT include this slash.

Key observations:

The same model works perfectly when set as the top-level "model" config
The same model works when selected via /models command
Only fails when specified in agent.plan.model, agent.build.model, or agent.explore.model
Setting "model": "" in agent config works (falls back to top-level model)

This appears to be a validation bug where the model name string is being processed incorrectly for agent-level model configurations with custom OpenAI-compatible providers.

Expected behavior

Agent should use the configured model my-custom-model without appending a trailing slash.

Actual behavior

Model validation fails because a / is appended to the model name, making it my-custom-model/ which doesn't match any configured model.

Workaround

Set "model": "" in agent configuration to fallback to the top-level model:

"agent": {
  "plan": {
    "mode": "subagent",
    "model": "",  // fallback to top-level model
    ...
  }
}




### Plugins

_No response_

### OpenCode version

1.1.21

### Steps to reproduce

1. Configure a custom provider with models in `opencode.json`:
```json
{
  "$schema": "https://opencode.ai/config.json",
  "model": "my-custom-model",
  "provider": {
    "my-provider": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "my-provider",
      "options": {
        "baseURL": "https://my-api-endpoint.com/v1"
      },
      "models": {
        "my-custom-model": {
          "name": "my-custom-model",
          "options": {
            "contextLength": 262144,
            "maxTokens": 16384
          }
        }
      }
    }
  },
  "agent": {
    "plan": {
      "mode":primary",
      "model": "my-custom-model",
      "temperature": 0.2
    }
  }
}

Open OpenCode
Switch to plan agent (or any agent with explicit model config)
See error: Agent plan's configured model my-custom-model/ is not valid

Screenshot and/or share link

No response

Operating System

Linux

Terminal

vscode terminal zsh

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

Ollama custom provider error : agent configured model is not valid #8950: Ollama custom provider error - agent configured model validation fails with custom providers
Amazon Bedrock Model in Agents not work #9603: Amazon Bedrock Model in Agents - agent model validation error with custom provider
Amazon Bedrock model in agent file #9392: Amazon Bedrock model in agent file - agent-level model validation failure
Server API ignores agent's configured model when only agent parameter is passed #7099: Server API ignores agent's configured model - agent-level model configuration not recognized
ProviderModelNotFoundError for custom provider in GitHub Action (worked until 07 Jan) #7958: ProviderModelNotFoundError for custom provider in GitHub Action - custom provider model resolution issues
Subagent with specific model results in model change in build and plan mode #6636: Subagent with specific model - agent-level model configuration handling issue
Agent variant incorrectly applied to all models, causing errors with OpenAI-compatible providers #11416: Agent variant incorrectly applied to all models with OpenAI-compatible providers (CLOSED)

These issues share similar patterns with agent-level model configuration failures, custom OpenAI-compatible provider handling, and model validation errors. Feel free to ignore if your specific case with the trailing slash appended to the model name is unique.

run opencode models to see list of available models. It looks like you aren't properly configuring your model id. It needs a provider prefix:

ex:

openai/gpt-5.2

This issue is real! Been hours with chatgpt and gemini to try to fix it but no luck at all.

Using latest opencode 1.2.15

Latest opencode.json created by gemini I'm using to test:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "local-ollama-host": {
      "api": "openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama"
      }
    }
  },
  "default_agent": "architect",
  "agent": {
    "architect": {
      "mode": "primary",
      "provider": "local-ollama-host",
      "model": "qwen-next-joomla:latest",
      "tools": {
        "write": false,
        "edit": false,
        "bash": false
      }
    }
  }
}

Wit it I get Agent architect's configured model qwen-next-joomla:latest/ is not valid

ollama list

NAME                       ID              SIZE     MODIFIED    
qwen-next-joomla:latest    29c27e42ac0e    51 GB    2 hours ago    
qwen3-coder-next:q4_K_M    ca06e9e4087c    51 GB    5 hours ago

To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.

Navigation Menu

Repository navigation

Agent model validation fails with trailing slash for custom OpenAI-compatible provider models #12062

Description

Description

Expected behavior

Actual behavior

Workaround

Screenshot and/or share link

Operating System

Terminal

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Agent model validation fails with trailing slash for custom OpenAI-compatible provider models #12062

Description

Description

Description

Expected behavior

Actual behavior

Workaround

Screenshot and/or share link

Operating System

Terminal

Activity

github-actions commented on Feb 4, 2026

rekram1-node commented on Feb 4, 2026

mgscreativa commented on Feb 27, 2026

github-actions commented on Apr 29, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions