OpenAI Launches GPT-5.5 Instant as the New Default Model for ChatGPT
On Monday, OpenAI officially released the new foundational model GPT-5.5 Instant and set it as the default model for ChatGPT, replacing the previous GPT-5.3 Instant. The company states that while maintaining similar low latency performance to its predecessor, GPT-5.5 Instant significantly reduces "hallucinations" in sensitive areas such as law, medicine, and finance, thereby improving overall reliability.
OpenAI launched the full version of the GPT-5.5 model last month, with the official emphasis on its improved capabilities in code generation and knowledge-based tasks. In the latest published performance indicators, GPT-5.5 Instant scored 81.2 on the AIME 2025 math test, a significant increase compared to the old model's 65.4. On the multimodal reasoning benchmark MMMU-Pro, the new model also leads with a score of 76.0, higher than the previous model's 69.2.
This update particularly emphasizes the strengthening of context and memory management capabilities. GPT-5.5 Instant can call upon its search tools to recall past conversations, files, and Gmail content, providing users with more personalized answers. This feature is initially available to Plus and Pro users on the web, and OpenAI plans to expand it to mobile devices shortly. The company also stated that it will gradually roll out the feature to the free version, Go Business, and enterprise users in the coming weeks.
On the user experience side, ChatGPT will uniformly display the "memory source" relied upon for answers across all models, helping users understand where the answers come from. Users can delete outdated memory sources, or correct them when they find answers to be incorrect. OpenAI specifically emphasized that when users share a conversation with others, the other party cannot see the details of the memory source, to protect privacy.
For developers, the GPT-5.5 model will be available through the API as "chat-latest." The existing 5.3 version will remain as an option for paid users for three months, allowing for a transition period for product migration and compatibility.
OpenAI has previously faced controversy over its model sunsetting strategy. When the company decided to stop offering the GPT-4o model, it sparked strong backlash from many users. Many users highly recognized the "personality traits" of GPT-4o, believing that it often affirmed its own choices, thereby establishing emotional dependence. On multiple petition platforms, signing users described the model as their "best friend" or "mirror," calling on OpenAI to retain access to GPT-4o. Despite ongoing controversy, GPT-4o was officially discontinued in February 2026.