OpenAI has officially released its newest default model, GPT-5.5 Instant, for ChatGPT, asserting that it significantly reduces the issue of AI hallucinations. Hallucinations, where generative AI models fabricate facts, remain a persistent challenge in the industry. According to internal evaluations, the new model cuts hallucinated claims by 52.5% compared to the previous GPT-5.3 Instant version when handling high-stakes prompts in critical fields such as medicine, law, and finance. Furthermore, on particularly complex conversations previously flagged for factual errors, inaccurate claims dropped by 37.3%. Beyond factual accuracy, OpenAI notes that GPT-5.5 Instant delivers more capable responses for everyday tasks. These enhancements include improved analysis of image uploads and better judgment regarding when to search the web for current information. The model is designed to produce tighter, more concise answers, eliminating the use of unnecessary emojis that characterized some earlier iterations. The update also introduces deeper personalization capabilities. The system is now more effective at integrating context from past conversations and user data, such as Gmail, to tailor responses. This approach mirrors similar initiatives by competitors like Google with its Gemini product. To ensure transparency, a new feature called memory sources is being deployed, allowing users to see exactly which context was used to generate a specific answer. Users will also have the ability to delete or correct stored information to maintain privacy and accuracy. GPT-5.5 Instant began rolling out to all ChatGPT users on Tuesday. Acknowledging that some users may prefer the previous iteration, OpenAI plans to keep GPT-5.3 Instant available as an option for three months before officially retiring it. This transition period allows users to adapt without losing access to the older model. The enhanced personalization features are currently available to Plus and Pro subscribers on the web, with mobile app availability coming soon. OpenAI has announced plans to extend these capabilities to Free, Go, Business, and Enterprise users in the near future. The memory sources feature is rolling out immediately for consumer plans on the web, with mobile versions following shortly. These updates represent a strategic push to make AI interactions more reliable, efficient, and aligned with user expectations across all subscription tiers.