Opencode for some reason does not know how to parse model data if used with local llama.cpp server. This jq filter parses the output of Open AI `v1/models` in a format that can be pasted into opencode config providers/llama.cpp models setting

parse-oai-modeldata.sh

#!/usr/bin/env bash
LLAMA_CPP_BASE_URL=https://llamacpp.your-url.com
curl -s ${LLAMA_CPP_BASE_URL}/v1/models | jq '[.data[] | 
  .status.args as $args | 
  {
    (.id): {
      name: .id,
      limit: (
        ($args | index("--ctx-size")) as $idx |
        if $idx then {context: ($args[$idx + 1] | tonumber), output: ($args[$idx + 1] | tonumber)} else empty end
      )
    }
  }
] | add'

The output will look like this:

{
  "GLM-4.7-Flash": {
    "name": "GLM-4.7-Flash",
    "limit": {
      "context": 184320,
      "output": 184320
    }
  },
  "JoyAI-LLM-Flash": {
    "name": "JoyAI-LLM-Flash",
    "limit": {
      "context": 128000,
      "output": 128000
    }
  },
  "LFM2-24B-A2B": {
    "name": "LFM2-24B-A2B",
    "limit": {
      "context": 32768,
      "output": 32768
    }
  },
....

So in your ~/.config/opencode/opencode.json you can have this snippet:

...
 "provider": {
    "llama.cpp": {
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://llamacpp.your-url-here.com/v1",
        "apiKey": "none",
        "api": "openai-completions"
      },
      "models": {
        "GLM-4.7-Flash": {
          "name": "GLM-4.7-Flash",
          "limit": {
            "context": 184320,
            "output": 184320
          }
        },
        "JoyAI-LLM-Flash": {
          "name": "JoyAI-LLM-Flash",
          "limit": {
            "context": 128000,
            "output": 128000
          }
        },
        "LFM2-24B-A2B": {
          "name": "LFM2-24B-A2B",
          "limit": {
            "context": 32768,
            "output": 32768
          }
        },
...