What can I to do ?

[MarkusAureus]

[Omegastick]

I'm guessing you need to use a model that supports tool calls.

[MarkusAureus]

Nice limiting...

[MarkusAureus]

This is supported..
OpenCode is nice sh1t.

[adamdotdevin]

this error is coming straight out of ollama, is there config you need to change on the ollama side?

[adamdotdevin]

ollama/ollama#8517 is it this issue?

[adamdotdevin]

maybe try another non-ollama local provider?

[jaswdr]

This is supported.. OpenCode is nice sh1t.

This is not particularly a limitation of OpenCode. The same error will happen if you use this model with any other tool. I suggest you to check the models that support tools and try it out. As example, most of the qwen models support it, like qwen3-8b. I've tested it by installing LLM Studio locally and it worked really well. Of course, due to the size of the model, specially when comparing with Claude 4 Sonnet, the responses are not perfect. But I recommend you to keep testing anyway, eventually you will find a model that fits your need, this is just the nature of running models locally.

[perfectra1n]

deepseek-r1:8b, llama2, and DeepSeek-Coder don't support tool calling, as the error states @MarkusAureus 😂😂😂

ollama/ollama#8517

Try qwen3:1.7b lol

[MarkusAureus]

I solved it.
I got it.
pip install open-interpreter