OpenAI to launch GPT-5, its most powerful model today: What to expect

OpenAI appears to be gearing up to unveil GPT-5, its next major AI model, during a livestream event this Thursday at 10AM PT. The teaser—“LIVE5TREAM”—all but confirms it, swapping the “s” in “livestream” for a “5.” Recent posts from OpenAI execs have only added fuel to the fire: CEO Sam Altman shared a screenshot with “ChatGPT 5” in the corner, while the head of applied research said he was “excited to see how the public receives GPT-5!”

What to expect

While OpenAI hasn’t revealed specifics, GPT-5 is widely expected to be a significant upgrade over GPT-4 and GPT-4o. Improvements could include better reasoning, longer context windows, fewer hallucinations, and improved reliability across domains like coding, maths, and scientific writing. Given OpenAI’s growing focus on AI agents, GPT-5 may also be more “tool-aware” and capable of complex multi-step tasks.

One major question is whether GPT-5 will be multimodal. While GPT-4o unified text, image, audio, and video processing in a single model, GPT-5 could push further by enabling more natural conversation and broader sensory input. Enhanced memory features may also be on the table, allowing it to retain context across sessions.