OpenAI Rolls Out GPT-5.5 Instant as ChatGPT Default
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI launched GPT-5.5 Instant as ChatGPT's default model on May 5, 2026, replacing GPT-5.3 Instant with what the company describes as smarter, more accurate, and more concise responses in a warmer, more natural tone
- The model is roughly half as likely to mislead users on medical and financial questions, per Digital Trends, with benchmark scores jumping from 78.5% to 85.6% on PhD-level science and 65.4% to 81.2% on competition math
- OpenAI also upgraded memory and personalization, letting ChatGPT draw context from saved memories, past chats, files, and connected Gmail accounts, with new 'memory sources' UI showing which context informed each response
- Sam Altman promoted the launch on X, arguing that the combination of speed, intelligence, personality, and memory improvements 'feels like a more-than-sum-of-the-parts' upgrade
- 9to5Mac and PCWorld highlighted a smaller user-experience win: the new model tones down the gratuitous emojis ChatGPT had been over-using in replies
- OpenAI engineer Michelle Pokrass said the team focused on 'factuality, crushing hacks, and improving the baseline intelligence,' framing 5.5 as a significant step on all three
Why it matters: For free-tier ChatGPT users—still the vast majority of the product's base—Thinking mode is unavailable, meaning GPT-5.5 Instant is the only model they touch, so the near-halving of misleading medical and financial answers lands directly on everyday health and money questions. With OpenAI pushing default upgrades roughly every model generation, the bar for what 'good enough' consumer AI looks like keeps moving.
Ask SkimNews
