What Is Mid-turn Steering on an OpenAI Model?

Nikunj Handa says mid-turn steering on GPT-6 lets someone inject messages while the model is still reasoning, and new instructions can go in as soon as a tool call finishes. The behavior is partly a model alignment capability, so the model has to be trained for it. Steering during reasoning was already possible in the app. The earlier version was not the best, and it has gotten much better.

With every new model, the team works closely with the post-training team and the research team to learn what is new, then exposes the new capabilities in the API. GPT-6 also launched async function calling, also called async tool calling. Tool calls take so long that model execution no longer has to pause while a tool runs. A tool call can start, the model keeps reasoning, and the result gets checked later.

The main goal in the API is to add a feature only after it is trained into the harness and good enough. Much of the new flow runs on WebSockets. He is unsure of the exact timing and places the launch a few months earlier. WebSockets open bidirectional communication with the model. On GPT-6 the channel carries async tool calling, async reasoning, and injected messages.