OpenAI is implementing immediate improvements, including resetting usage limits twice and adjusting default settings to prevent unnecessarily expensive compute usage for its GPT 5.6 Sol model and agentic features. A larger set of improvements will land next week, bringing chats and projects back to the sidebar and clarifying the distinct use cases for ChatGPT Work and Codex.
OpenAI is rapidly iterating on its new model and agent platform, indicating a willingness to adjust product strategy based on user experience and compute consumption patterns.