Release notes

For new Health plugin connections, ChatGPT will now default to each user’s global plugins permissions setting.

If you make no changes to your Plugins settings, the default permission setting is “Allow low-risk actions.” This means that once you’ve connected your information in Health, ChatGPT can use it when relevant without asking for permission each time, unless you are attempting a sensitive action, such as asking ChatGPT to email a training plan based on your data to a running partner.

This change reflects what we’ve seen from users: more than 70% of Health users already choose to allow ChatGPT to use data connected in Health without asking for permission each time.

You can change your Health permissions anytime in Settings → Plugins → Health.

We’re retiring automatic switching from Instant to Thinking (reasoning) for ChatGPT Plus and Pro users globally. You can still select an available option in the model picker to give ChatGPT more time to think or reason.

We’re also removing the Higher intelligence setting from ChatGPT on the web for these plans. ChatGPT can still switch automatically for safety purposes.

On October 14, 2026, GPT‑5.5 will retire from ChatGPT, ChatGPT Work, and Codex on all plans, including consumer, Business, Enterprise, and Edu plans. This retirement does not apply to the OpenAI API.

For Codex with ChatGPT sign-in, switch from gpt-5.5 to gpt-5.6-sol (GPT‑5.6 Sol) before the retirement date.

Update workspace defaults, saved model settings, managed configurations, custom agents, scheduled tasks, and scripts that still select gpt-5.5.

ChatGPT Business admins can use Model Test to understand which models a member can access and which settings contribute to that access. Open Admin Console, select your ChatGPT workspace, then use Models > Test to search for a member by name or email. Testing does not grant model access, change permissions or usage limits, or override the member’s seat type or workspace plan.

Changes made through the Codex Policies & Configurations interface are recorded in workspace audit logs. Authorized admins can review the logs in the UI or retrieve them through the Compliance API.