Realtime AI News
New GPT Image 2.5 surfaces inside ChatGPT before launch, can already fake GPT-6 event images
OpenAI's new image model GPT Image 2.5 is already surfacing inside ChatGPT before an official announcement, with test results showing it can fabricate convincing GPT-6 launch photos and tweet screenshots. Compared with GPT Image 2, it generates faster, looks more photorealistic, keeps faces and embedded text stable, and fixes the earlier noise artifacts.
Just as OpenAI wrapped up its GPT-6 launch, a batch of highly convincing “photos from the event” began circulating on social networks. Chinese tech outlet QbitAI reports they were produced by OpenAI's new image model, GPT Image 2.5, which has not been officially released but has already reached some users as a new popup prompt inside ChatGPT.
The version number is the first surprise: instead of the GPT Image 2.1 many expected, the model jumped straight to 2.5, suggesting a substantial upgrade. According to test results circulating online, it can now fabricate official-looking GPT-6 launch images as well as an all-staff letter and tweet screenshots from Sam Altman, with text that shows almost no flaws apart from an incorrect avatar.
On the LMArena leaderboard, an anonymous model codenamed luna-lisa-alpha is widely believed to be the new GPT Image. Community testing has identified several traits: very fast generation, more lifelike photos, faces that stay stable across multiple images of the same character, embedded text that resists distortion, and a fix for the noise artifacts that plagued GPT Image 2.
One tricky test asked the model to produce a realistic phone photo of a handwritten article in Albert Einstein's handwriting on ruled paper, with a coffee stain in the upper-right corner. Placed next to the real Einstein handwriting, the output is nearly indistinguishable.
OpenAI's product placement hints at how much weight image generation now carries: the ChatGPT web app briefly moved image generation up to the second slot in its main menu, right below “new chat,” before hiding it again. The rising prominence of image generation inside ChatGPT suggests it will become one of the flagship features of the GPT-6 era.
The open question is when OpenAI will make the model official, since the final version number, rollout scope, and API pricing have not been announced. How image generation will combine with GPT-6 Astra's push toward “completing work” is the next thing to watch.
Why it matters
GPT Image 2.5 shows OpenAI's image generation stack improving quickly while still unofficial, signaling image generation as a core GPT-6-era feature and adding fresh urgency to the fight against AI-generated fakes.
Nearby Updates
All09/04, 05:57
Tenable and OpenAI Are Building a Way to Check 100+ Community-Built AI Tools
According to Stock Titan, more than 100 community-built AI tools have been listed, and cybersecurity firm Tenable is working with OpenAI on a way to check them. The partnership brings security vetting into the fast-growing ecosystem of community AI tools, though the mechanism's details and timeline have yet to be disclosed.
09/04, 05:50
OpenAI Officially Releases GPT-6: 'Welcome to the AGI Era'
OpenAI has officially released GPT-6, its new flagship model, framing the launch with a direct "welcome to the AGI era" message. Chinese tech outlet QbitAI reported the release as a breaking event, and the announcement now raises the question of what OpenAI will show to back the claim.
09/04, 05:43
OpenAI commits $1 billion to cyberdefense efforts amid AI safety scrutiny
OpenAI has committed $1 billion to cyberdefense efforts, Reuters reports, in a move that arrives amid heightened scrutiny of the company's approach to AI safety. The sizable pledge signals that OpenAI is putting concrete resources behind its security posture rather than relying on statements alone.
09/04, 05:02
OpenAI's cybersecurity AI planned for Elastic Security, no separate model infrastructure needed
OpenAI's cybersecurity AI is planned for integration into Elastic Security, with no separate model infrastructure required for customers, according to an announcement carried by Stock Titan. If it ships as described, security teams would get the capability directly inside Elastic's existing platform at a much lower adoption cost.