The updated default model of ChatGPT is now 50% less likely to provide misleading information regarding medical and financial inquiries.

      GPT-5.5 Instant is now the default model for ChatGPT, promising notable reductions in hallucinations in areas such as medicine, law, and finance, along with shorter responses and improved personalization.

      OpenAI has replaced GPT-5.3 Instant with GPT-5.5 Instant as the standard ChatGPT model, with the update being made available to all users starting today. The updated model aims to enhance three main aspects: accuracy, conciseness, and personalization.

      Fewer hallucinations, less unnecessary content

      According to OpenAI, the improvements in accuracy are particularly significant in critical subjects. Internal tests showed that GPT-5.5 Instant produced 52.5% fewer hallucinated claims compared to its predecessor on prompts related to medicine, law, and finance, and 37.3% fewer inaccurate claims on conversations that users had previously reported as erroneous.

      GPT-5.5 Instant is being implemented in ChatGPT. This major upgrade offers you smarter, clearer, and more personalized responses in a friendlier, more natural tone. Additionally, it is more concise, responding in fewer words, which we noted you desired. We believe you will enjoy interacting with it.

      The new model also demonstrates improved capabilities in photo analysis, STEM inquiries, and web search decisions. Responses are shorter and streamlined, with OpenAI stating that the new model utilizes 30.2% fewer words and 29.2% fewer lines than GPT-5.3 Instant.

      Users will also find fewer unnecessary follow-up questions and enhancements in formatting.

      Enhanced memory and user control

      Furthermore, GPT-5.5 Instant is better at incorporating relevant context from past conversations, uploaded documents, and connected Gmail accounts to tailor responses without needing users to repeat information.

      In conjunction with this, OpenAI is launching "memory sources" across all ChatGPT models. This new feature will inform users about the context that influenced a response, allowing them to delete or amend any part of it directly from settings.

      The rollout of GPT-5.5 Instant will extend over the next two days to all ChatGPT users. GPT-5.3 Instant will remain accessible to paid users for three months through model configuration settings before being phased out. Enhanced personalization will initially be available to Plus and Pro users on the web, with Free, Go, Business, and Enterprise users to follow in the coming weeks.

      Memory sources are being introduced across all ChatGPT consumer plans online, and OpenAI indicates that improvements in personalization and memory sources will soon be available on mobile as well.

      Pranob is an experienced tech journalist with over eight years of experience in covering consumer technology. His writings have been…

      Your ChatGPT history serves as an unseen personality test

      Researchers have trained an AI to deduce personality traits based on your ChatGPT interactions, and its accuracy was impressively high.

      Each time you ask ChatGPT to draft an email, discuss a relationship issue, or search for symptoms, you may be sharing more than just a request. As reported by TechXplore, researchers at ETH Zurich developed an AI model capable of predicting personality traits directly from actual ChatGPT conversation logs, and it accurately identified personality traits.

      The study analyzed 62,090 real conversations from 668 ChatGPT users. Participants also took a standardized personality assessment, giving researchers a reference point for evaluation. The AI was trained to categorize each user as low, medium, or high across five personality traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism.

      Read more

      iOS 27 may allow users to select their preferred AI model for text and image tasks

      Apple is not racing to create the most advanced AI; instead, it is establishing the best platform for others to excel, and iOS 27 could mark the beginning of a very competitive environment.

      There will be a significant transformation in AI functionality across Apple's devices, with users at the forefront. According to Bloomberg’s Mark Gurman, iOS 27, along with its major counterparts (iPadOS 27 and macOS 27), will enable users to choose among various third-party AI models for tasks within Apple’s core intelligence features.

      You can compare it to using your TV remote to switch channels; here, the remote is iOS 27, and the channels are popular AI models you may have used in other contexts.

      Read more

      32GB RAM for Windows 11? It sounds like a problem on Microsoft’s end!

      If your operating system requires more memory, perhaps the issue lies not with the user's PC.

      There was a time when purchasing a PC felt... logical. 8GB of RAM was sufficient, 16GB felt like a statement, and anything beyond that was typically for those engaged in genuinely demanding tasks. This equilibrium existed because software was designed with hardware limitations in mind. Today, that balance has quietly fallen apart, and Microsoft appears to accept this change.

      The company's recently retracted guidance, casually suggesting 16GB as the minimum and 32GB