OpenAI has updated its Codex platform, enabling the million-token context window feature of the GPT-5.6Sol model for all ChatGPT accounts. Previously, this capability was only accessible via API keys. Users with ChatGPT Plus and Pro subscriptions can now manually activate it in Codex to handle more code, tool outputs, and conversation records before auto compressing older data. This improves efficiency for tasks like large codebase refactoring and complex project debugging. The team advises developers to set the limit based on specific needs to avoid rapid quota consumption, as the default setting balances performance and cost.