GPT-5 Launch Might Be Imminent: More Leaks Detail Some Of The New AI's Main Features

By Chris Smith
AI apps on an iPhone display, including ChatGPT. miss.cabul/Shutterstock

Reports last week said ChatGPT's big GPT-5 upgrade would launch in August, but didn't reveal a release date. The best estimate was "early August," though OpenAI can always postpone AI model releases. The open-source ChatGPT release, which has been delayed more than once already, should also launch before GPT-5 rolls out, according to last week's reports.

What's certain is that GPT-5 is coming. Sam Altman said long ago that the next-gen ChatGPT model would be available this year, teasing a new unified AI experience, in which the chatbot would be able to automatically choose the right tools depending on a user's prompt.

Last week, Altman was on tour in Washington, D.C., and he further teased what he'd seen while testing GPT-5, suggesting an even more powerful AI is about to be available to users. The executive also revealed he aims to give everyone GPT-5. Leaks suggested that premium ChatGPT users would get better features and more access than ChatGPT Free subscribers, though.

While OpenAI isn't ready to tease the GPT-5 launch like it did with ChatGPT Agent ahead of the AI agent's launch, we're seeing more signs that GPT-5's arrival is imminent. Some testers are already using the new models, while others are sharing leaks about it.

GPT-5 to launch next week?

Concept image for ChatGPT o3 reasoning mode. OpenAI

Menlo Ventures AI investor Deedy Das posted the following tweet on X, claiming OpenAI will launch GPT-5 next week. According to information from an anonymous source that provided "reasonable proof," GPT-5 will support context windows of up to 1 million tokens (input) and up to 100,000 tokens (output). GPT-5 will also support the Model Context Protocol (MCP) standard, which lets AI models connect to external data sources and "parallel tool calls." Das also said GPT-5 will support dynamic (short and long) reasoning, Code Interpreter, and other tools.

The leaker also mentioned the codenames OpenAI is supposedly using for GPT-5 models: "o3-alpha > nectarine (GPT-5) > lobster (mini) > starfish (nano)." Put differently, Nectarine, Lobster, and Starfish are the three GPT-5 models that will supposedly be available to users.

These codenames also appeared in a report from The Neuron earlier this week, which speculated that GPT-5 is being tested. "At least six anonymous models—Zenith, Summit, Lobster, Nectarine, Starfish, and o3-alpha—that are supposedly outperforming nearly every other known model," have been spotted on LMArena. This report indicates Zenith might be the top-tier version of GPT-5, not Nectarine.

Separately, X user chatgpt21 posted screenshots supposedly coming from the head of design at Cursor, who has been testing GPT-5 Alpha. The images in the following tweet show an interaction with an AI model in a Mac app, probably Cursor. The model name is blurred but in a way that's still legible. It reads gpt-5-alpha, likely a test AI model that some people have been able to use for a while. Finally, some testers have noticed a GPT-5-powered Smart Mode that Microsoft is preparing for Copilot.

GPT-5 teasers

ChatGPT Agent UI for ChatGPT Plus users. Chris Smith, BGR

As a longtime ChatGPT user, I'm curious to see what GPT-5 can do to improve my daily interactions with the chatbot. I go to ChatGPT with all sorts of questions throughout the day, with o3 being the main model I use. I like its ability to take its time to provide answers, though o3 makes mistakes too. I'm also using the AI to craft training sessions for my marathon races and deal with all my travel needs. I started giving ChatGPT Agent complex tasks to see what it can do, though I'm only testing the AI agent for now. What I'm getting at is that I'm already well aware of what the AI can do for me. Hearing Sam Altman's GPT-5 teasers last week only made me more interested in it.

"This morning, I was testing our new model, and I got a question. I got emailed a question that I didn't quite understand. And I put it in the model, this is GPT-5, and it answered it perfectly," Altman said during a podcast interview. "And I really kind of sat back in my chair, and I was just like, 'Oh man, here it is moment' [...] I felt like useless relative to the AI in this thing that I felt like I should have been able to do, and I couldn't. It was really hard. But the AI just did it like that. It was a weird feeling."

I'm not sure if GPT-5 will impress me similarly, considering I might not use the AI for the same complex purposes. But I do want a fast, universal AI model that knows what tools to select, when to reason, and when to provide a Deep Research report without me providing specific instructions.

Recommended