Hi, what is everyone’s experience with the new update just rolled out today for users using 4o? Are the rules a lot tighter now for creative writing? here’s the release notes >>> We are replacing the existing GPT‑4o-based model for Operator with a version based on OpenAI o3. The API version will remain based on 4o.
o3 Operator uses the same multi-layered approach to safety that we used for the 4o version of Operator and described in our original Operator System Card(opens in a new window). Compared with other models in the o3 family, o3 Operator was fine-tuned with additional safety data for computer use, including safety datasets designed to teach the model our decision boundaries on confirmations and refusals.