OpenAI Adds Auto, Fast and Thinking Modes to GPT-5 ChatGPT

OpenAI has expanded ChatGPT’s functionality by introducing three operating modes for its GPT-5 model—Auto, Fast and Thinking—giving users more control over performance and speed. The company recommends Auto for general use, while Fast and Thinking provide alternative trade-offs for latency and depth of reasoning. The update lifts rate limits to 3,000 messages a week when GPT-5 runs in Thinking mode and allows additional capacity beyond that threshold. OpenAI also disclosed a 196,000-token context window for GPT-5 Thinking and reinstated GPT-4o in the model picker for all paying customers, broadening options for enterprise and individual subscribers.