OpenAI has launched a free, specialized ChatGPT version for US healthcare professionals, featuring clinical workflows, real-time literature citations, and optional HIPAA compliance.
Starbucks has launched a beta mini-app inside ChatGPT that uses AI to recommend drinks based on your text descriptions or uploaded photos, with ordering integrated directly in the chat.
OpenAI's next big model, likely called GPT-5.5, has surfaced in the Codex developer interface, but it returns an error when selected. The company hasn't officially announced it.
Anthropic removed Claude Code from its Pro subscription in a quiet test, pushing new users to a $100/month tier. OpenAI immediately countered, pledging to keep its Codex coding assistant free.
OpenAI's ChatGPT Images 2.0 can now generate 2K images, offers a 'Thinking' mode for consistency, and supports conversational editing.
Leaked screenshots show OpenAI is developing a "Hermes" agent builder for ChatGPT, letting users create scheduled assistants that connect to Slack, apps, and files.
A leaked presentation shows ad tech firm StackAdapt is selling ChatGPT ad slots for $15–60 CPM with a $50k minimum spend, as OpenAI's ad push sparks employee departures and mixed reactions.
OpenAI is preparing to add customizable pixel-art avatars to its Codex developer tool, letting users pick or create visual companions that chat via speech bubbles during coding sessions.
A developer claims a logic flaw in ChatGPT's iOS subscription system could allow one legitimate Apple Pay receipt to activate Plus for multiple accounts, but the report lacks proof and hasn't been verified.
The NSA has begun using Anthropic's powerful new Mythos Preview AI for cybersecurity, even as the Pentagon debates the risks, revealing a government split over AI adoption.
