Model not returning the answers.

[Regboy744]

Question

Description
When using qwen2.5-coder:14b with OpenCode, the model attempts to call tools (like todorantum) for simple conversational inputs instead of responding normally.
Steps to Reproduce

Set up OpenCode with Ollama provider
Use model: qwen2.5-coder:14b
Send a simple greeting: hi
Model responds with a tool call instead of a direct response

Expected Behavior
User: hi
Assistant: Hello! How can I help you today?
Actual Behavior
User: hi
Assistant: {
"name": "todorantum",
"arguments": {
"input": "hi"
}
}
Environment

Provider: Ollama
Model: qwen2.5-coder:14b
Ollama Version: 0.13.5
GPU: NVIDIA RTX 5070 Ti (16GB VRAM)
OS: Arch Linux

Model
architecture qwen2
parameters 14.8B
context length 32768
embedding length 5120
quantization Q4_K_M

Capabilities
completion
tools
insert

System
You are Qwen, created by Alibaba Cloud. You are a helpful assistant.

License
Apache License
Version 2.0, January 2004
...

"provider": {
"ollama": {
"npm": "@ai-sdk/openai-compatible",
"name": "Ollama (local)",
"options": {
"baseURL": "http://localhost:11434/v1"
},
"models": {
"qwen2.5-coder:14b": {
"name": "qwen2.5-coder:14b",
"tools": true,
"thinking": true,
}
},
}
}

[github-actions]

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

• Local Ollama models are not agentic #5694: Local Ollama models are not agentic - Similar problem with Ollama models not functioning correctly
• Cannot use tools with qwen2.5-coder and Ollama #2728: Cannot use tools with qwen2.5-coder and Ollama - Directly related to the qwen2.5-coder model showing tool call JSON instead of executing
• Local Ollama tool calling either not calling or failing outright #1034: Local Ollama tool calling either not calling or failing outright - Similar pattern where Ollama models generate tool calls but don't execute them properly

Feel free to ignore if your specific case is different from these.

[rekram1-node]

what is the num ctx u set for the model?