Levy

Hey Great blog post! I noticed you published the models tested for the Gaia/Math benchmarking—thanks for sharing that.

I was wondering if you could let me know which server you used to run the models? I’m currently using vLLM, but it seems like it doesn’t yet support tool_choice="required". To make it work, I had to tweak the model code to force it to return a tool for execution.

Thanks in advance for any insights!