OpenAI CEO Sam Altman recently announced significant updates to ChatGPT, allowing users greater control over the new GPT-5 model. In a post on X, he introduced three response modes: “Auto,” “Fast,” and “Thinking,” catering to different user needs for speed and detail. The GPT-5 Thinking mode features a weekly limit of 3,000 messages, with an option for additional capacity via GPT-5 Thinking mini. Enhanced with a 196k-token context limit, GPT-5 can manage longer conversations and documents, adapting based on usage. Altman confirmed the return of GPT-4o for paid users and introduced a “Show additional models” toggle for options like o3 and 4.1. The exclusive GPT-4.5 model remains for Pro subscribers due to high GPU costs. OpenAI is also updating GPT-5’s personality to be “warmer.” Altman characterized GPT-5 as a major upgrade, especially in coding and complex reasoning, enhancing ChatGPT’s expertise across various topics.
Source link
Share
Read more