ollama-cloud: reasoning variants not passing parameter to Ollama API #36638

Problem

The built-in ollama-cloud provider's reasoning variants for thinking models do not correctly control the thinking effort level. When switching between high and max variants, the Ollama Chat API request does not include the think parameter, which is how the Ollama API controls reasoning intensity per the spec.

Steps to Reproduce

1. Connect to Ollama Cloud via /connect
2. Select a thinking model, e.g. glm-5.2 (has the thinking tag on ollama.com)
3. Switch between high and max variants
4. The thinking content in the response shows no meaningful difference in reasoning depth between variants
5. The think parameter is absent from the API request sent to Ollama

Workaround

Manually override variants in opencode.json:

{
  "provider": {
    "ollama-cloud": {
      "models": {
        "glm-5.2": {
          "id": "glm-5.2:cloud",
          "reasoning": true,
          "variants": {
            "high": { "think": "high" },
            "max":  { "think": "max" }
          }
        }
      }
    }
  }
}

Expected Behavior

The built-in reasoning variants for ollama-cloud should correctly pass the think parameter as a top-level field in the Ollama Chat API request, matching the Ollama API spec.

References

• Ollama Chat API: POST /api/chat — think is a top-level parameter accepting "high", "medium", "low", "max" or boolean
• GLM 5.2 on Ollama: https://ollama.com/library/glm-5.2 (supports effort level control: High and Max)