OpenAI releases GPT-5.5 Instant as the new default ChatGPT model, replacing GPT-5.3 Instant with improvements in math reasoning and multimodal tasks.

GPT-5.5 Instant becomes the default ChatGPT model on Tuesday, succeeding GPT-5.3 Instant. The model achieves a score of 81.2 on the AIME 2025 math test, compared to 65.4 for the older model, and scores 76 on the MMMU-Pro multimodal reasoning benchmark versus 69.2 for its predecessor.

The model adds context management via a search tool that can refer to past conversations, files, and Gmail to provide personalized answers. Memory sources are now shown across all models, and users can delete or correct outdated sources; shared chats do not expose memory sources to recipients.

For developers, GPT-5.5 will be available through API as 'chat-latest,' while GPT-5.3 remains available for paid users for only three months. The search feature rolls out first to Plus and Pro users on the web, with plans to extend to Free, Go Business, and enterprise users in the coming weeks.