Hi everyone,
Yesterday, OpenAI released two open-source models: gpt-oss-120b and gpt-oss-20b.
I’m wondering how I can integrate them as chat models for my AI agent.
Has anyone tried this yet?
I’ve tried dropping it in to my ai agent workflow, but always get the following error when trying to access the gpt-oss:20b through my model selector.
——
Problem in node ‘Ai Agent
template: :108:130: executing “” at : error calling index: reflect: slice index out of range
——
The model is running via ollama on the server and ollama run cmd opens the chat interface just fine.
I also have this agent hooked up to about 20 tools (normally just let it talk to gpt-4.1) and didn’t change any of my system messaging, tool descriptions, chat model settings, or Ai Agent node settings. So not exactly sure what’s going wrong, but it sure isn’t a drop in replacement.
I might take a run at making a new agent from scratch with the different gpt-oss models and see how far it can go before it breaks.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.