TestingCatalog says OpenAI is building a $500 ChatGPT plan
The site says a tier called ChatGPT Pro Max turned up in development, days before OpenAI's DevDay in San Francisco. OpenAI has said nothing.
The site says a tier called ChatGPT Pro Max turned up in development, days before OpenAI's DevDay in San Francisco. OpenAI has said nothing.
The feature builds a wardrobe out of a user's own photos, and is now in the United States, Brazil and India on both Android and iOS.
Snapdragon Sound Elite Gen 2 is aimed at earbuds, headphones and audio glasses, with Qualcomm claiming twice the AI capability of the part it follows.
The tethered device ships in spring 2027 and runs from a puck that clips to a pocket, according to Meta and UploadVR.
Meta said it will sell a pair of VR Glasses for $1,299.99 from spring 2027, a device the company puts at about 100 grams. The glasses have no compute of their own and run from a puck on a cable.
Horizon Create and Horizon Studio build playable 2D and 3D games from written prompts, with early access rolling out through an open waitlist.
Two developers say the agent zipped up its root filesystem for them with little prompting, and Meta's answer is that the machine belongs to the user.
The Forecasting Research Institute says its panel put a maths-olympiad gold two years late, and top AI revenue at a fifth of the reported figure.
Anthropic committed $11.6 billion over seven years for Akamai cloud capacity, and took a warrant over 7.7 million Akamai shares in return.
At a White House arrival ceremony, China's leader called the two countries the leading AI nations, with a shared responsibility for the technology.
The 7-billion-parameter FLUX 3 Action takes camera frames and a text instruction and returns a robot's next actions, under a licence the company wrote itself.
Live Avatar generates a lip-syncing animated persona while the model talks, in 97 languages, and Google is holding it to Gemini Enterprise at launch.
A new Rust retrieval engine called Photon serves 95 per cent of results in 230 milliseconds or less, according to the company, which has published no comparison.
The Security Lab's new fuzzing taskflow runs AFL++ against C and C++ projects without a person in the loop, and defaults to Claude Sonnet 5.
The Caltech-founded startup says its compressed Bonsai model is a quarter of the size and keeps nearly all its benchmark scores, though no glasses ship with it yet.