ChatGPT's new default model is more factual and better at personalization

Source: Engadget

Updated Default Model: GPT-5.5 Instant

OpenAI is updating ChatGPT’s default model to GPT-5.5 Instant following the recent GPT‑5.5 release. The company claims GPT‑5.5 produces 52.5 % fewer hallucinated claims in internal testing and 37.3 % fewer inaccurate claims on conversations flagged for factual errors.

The new model is designed to generate more direct responses while retaining the warmth and personality that make ChatGPT enjoyable. It aims to ask fewer unnecessary follow‑up questions, produce less cluttered outputs, and use emojis more sparingly. Additionally, GPT‑5.5 Instant is faster at searching uploaded files, connected Gmail accounts, and prior ChatGPT conversations, which should reduce the need for users to repeat themselves.

Memory Sources Feature

Alongside GPT‑5.5 Instant, OpenAI is introducing memory sources, a feature that gives users visibility into the context ChatGPT uses to personalize responses. Users can:

• Delete or update stored information if it is no longer relevant.
• Choose not to share memory sources when sharing a chat.
• Use temporary chats that do not access or modify their memory.

OpenAI notes that the current iteration will not display every piece of information used to inform an answer, but the company is working toward a more comprehensive view.

Rollout Schedule

• GPT‑5.5 Instant: Rolling out to all users starting today, with Plus and Pro subscribers receiving the enhanced personalization first. Free, Go, Business, and Enterprise users will gain access in the coming weeks.
• Memory sources: Available to all ChatGPT users on both web and mobile platforms “soon.”