anomalyco/opencode
1
0
Fork 0
mirror of https://github.com/anomalyco/opencode.git synced 2026-07-22 02:05:26 -04:00
Code Issues 5.9k Packages Projects Releases 19 Wiki Activity

Local Ollama model returns non-human message #8897

New Issue
Open
opened 2026-02-16 18:11:07 -05:00 by yindo · 1 comment
yindo commented 2026-02-16 18:11:07 -05:00
Owner
Copy Link
Copy Source

Originally created by @samchoiml-cmd on GitHub (Feb 9, 2026).

Originally assigned to: @rekram1-node on GitHub.

Question

I have newly installed a WSL ubuntu with opencode on it.
Another machine (GPU installed) is running ollama with qwen2.5-coder:32b.

opencode.json as below

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ollama (local)",
      "options": {
        "baseURL": "http://192.168.0.66:11434/v1"
      },
      "models": {
        "qwen2.5-coder:32b": {
          "name": "qwen2.5-coder:32b"
        }
      }
    }
  }
}

when i communicate with opencode in terminal, it returns below structure

{"name",: "todowrite", "arguments": {"todos": {"content": ""check if there are any README or other documentation files to learn more about the project structure.", "status": "pending", "priority": "high", "id: "1"}, "content": "if no relevant documentation is found, create a basic README file for future reference.", "status": "pending", "priority": "low", "id":"2"}}}

Compared to GLM or Big Pickle, it seems that qwen2.5-coder:32b is not producing output in a human-readable format similar to those models.

Could this be a configuration issue?

Originally created by @samchoiml-cmd on GitHub (Feb 9, 2026). Originally assigned to: @rekram1-node on GitHub. ### Question I have newly installed a WSL ubuntu with opencode on it. Another machine (GPU installed) is running ollama with qwen2.5-coder:32b. opencode.json as below ``` { "$schema": "https://opencode.ai/config.json", "provider": { "ollama": { "npm": "@ai-sdk/openai-compatible", "name": "Ollama (local)", "options": { "baseURL": "http://192.168.0.66:11434/v1" }, "models": { "qwen2.5-coder:32b": { "name": "qwen2.5-coder:32b" } } } } } ``` when i communicate with opencode in terminal, it returns below structure `{"name",: "todowrite", "arguments": {"todos": {"content": ""check if there are any README or other documentation files to learn more about the project structure.", "status": "pending", "priority": "high", "id: "1"}, "content": "if no relevant documentation is found, create a basic README file for future reference.", "status": "pending", "priority": "low", "id":"2"}}}` Compared to GLM or Big Pickle, it seems that qwen2.5-coder:32b is not producing output in a human-readable format similar to those models. Could this be a configuration issue?
yindo added the windows label 2026-02-16 18:11:07 -05:00
yindo commented 2026-02-16 18:11:07 -05:00
Author
Owner
Copy Link
Copy Source

@github-actions[bot] commented on GitHub (Feb 9, 2026):

This issue might be related to existing model output formatting issues. Please check:

  • #4954: google/gemini-3-pro-preview output formatting went off the rails - similar issue with model output formatting degradation
  • #7779: GLM 4.7 thinking process not properly formatted - thinking content appearing as plain text instead of being formatted
  • #10456: schema-constrained structured outputs - feature request for enforcing JSON schema compliance in model output

The core problem appears to be model-specific output formatting issues where models don't produce human-readable output as expected. If qwen2.5-coder:32b is returning raw JSON instead of formatted messages, this could be related to how the model is configured or how its output is being parsed.

Feel free to check those issues to see if they provide any solutions or context for your specific case with Ollama.

@github-actions[bot] commented on GitHub (Feb 9, 2026): This issue might be related to existing model output formatting issues. Please check: - #4954: google/gemini-3-pro-preview output formatting went off the rails - similar issue with model output formatting degradation - #7779: GLM 4.7 thinking process not properly formatted - thinking content appearing as plain text instead of being formatted - #10456: schema-constrained structured outputs - feature request for enforcing JSON schema compliance in model output The core problem appears to be model-specific output formatting issues where models don't produce human-readable output as expected. If qwen2.5-coder:32b is returning raw JSON instead of formatted messages, this could be related to how the model is configured or how its output is being parsed. Feel free to check those issues to see if they provide any solutions or context for your specific case with Ollama.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
acp
bad
bug
discussion
docs
documentation
enhancement
extension
good first issue
growth
help-wanted
mcp
model-problem
needs:compliance
nix
opentui
perf
pull-request
Mirrored from GitHub Pull Request
 tui
web
windows
zen
No Label windows
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#8897
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?