OpenAI

Up for a challenge? We know good ideas can come from anywhere. So we’re hosting an open research challenge asking people to build the most efficient pretrained model under extreme constraints. Introducing: Parameter Golf. Details + submissions at https://lnkd.in/gmwiSU58

AI is starting to help solve real issues in healthcare for patients and doctors. OpenAI’s Head of Health Dr. Nate Gross, MD and Health AI Research Lead Karan Singhal join Andrew Mayne to discuss how we're building new models and products to meet the world's health needs. https://lnkd.in/dr9D6-8i

We must continue developing concrete ways for AI to help close longstanding access gaps in healthcare - here is an example based on work we supported at OpenAI: In one of the largest real-world studies of patient-facing clinical AI to date, Stanford, Healthvana, and collaborators published findings in Nature Digital Medicine this week. The researchers evaluated Healthvana’s AI-powered chatbot deployed across AIDS Healthcare Foundation clinics in 15 states to support HIV prevention. Of the 155,000+ patients, those who engaged with the OpenAI-based chatbot were 3x more likely to start PrEP (HIV prevention medication) than those who didn't. Follow-up care rates nearly doubled (57% vs. 32%), and appointment attendance jumped from 54% to 66%. Importantly, 80% of AI users came from racial and ethnic minority communities, which are disproportionately affected by HIV, and historically the hardest to reach through traditional healthcare channels. It’s clear that personalized support made this effective: ➡️ The AI tool communicated in 18 languages ➡️ It was available outside clinic hours (nearly 30% of messages were sent after hours) ➡️ It offered culturally resonant personas — 80% of users chose the LGBTQ+-affirming "Glitter Byte" option over the standard clinical tone ➡️ It reduced stigma as a barrier by giving patients a private, judgment-free space to ask sensitive questions And rather than replacing clinicians, 56% of clinic staff reported saving time on routine tasks, freeing them up for more complex care. The AI handled ~50% of inbound patient messages. Here’s to many more examples of this type of thoughtful, expert-driven AI implementation! 

Codex for non-technical people 👇 What have you been building with Codex?

Lauren Bannon, who runs a marketing business and raises two daughters, credits ChatGPT for saving her life after an illness that couldn’t be diagnosed. In February 2024, she started waking up with hands that did not work the way they always had. Her pinky fingers “weren’t bending in the morning and night.” Before long she was also experiencing heart palpitations and a reflux-like sensation when she leaned down. An X-ray came back normal. A rheumatology specialist labeled it seronegative rheumatoid arthritis and prescribed medication, yet Bannon hesitated. “I have never been sick,” she told her husband, and she felt she was too young for arthritis to make sense. She could not shake the sense that something was being missed. For many months, she tried to push through. But by September, she was hit with excruciating stomach pain, a relentless burning that left her surviving off toast and water. A gastrointestinal scope showed inflammation and she was put on proton pump inhibitors, but the inflammation just wouldn’t stop. At home, often late at night, Bannon turned to ChatGPT, a tool she already used for her work in marketing, even though her husband thought she was crazy for talking to a computer. She began entering her symptoms and blood results, looking for patterns across months of data that no single appointment could easily address. When she asked, “what mimics rheumatoid arthritis?” ChatGPT pointed her toward Hashimoto’s disease, an autoimmune thyroid condition where the body attacks itself. ChatGPT urged a specific next step: “Get your doctor to test your TPO level,” referring to a thyroid antibody. Her clinician initially pushed back, because Lauren’s other thyroid labs were normal, but armed with her knowledge and desperate for real answers, Bannon pushed harder. “I need to know what’s going on with me,” she told her doctor. “I want to test it.” The thyroid peroxidase antibody result came back elevated at 220, roughly six times normal levels, confirming an auto-immune condition like Hashimoto’s and leading her to an endocrinologist. From there, a neck scan revealed two lumps, and a biopsy confirmed they were cancer, including an aggressive mutation and probably related to the Hashimoto’s. Bannon then underwent a full thyroid removal, lymph-node surgery, and radioactive iodine to kill the remaining cancer cells, with ongoing scans to monitor whether it’s still in her neck. She believes ChatGPT intervened just in time to prevent the disease from spreading unchecked. “Without it, the cancer would have spread,” she said. If she had simply taken the rheumatoid arthritis medication and stopped searching, “God knows where it would have gone,” she said. ChatGPT empowered Lauren to advocate for herself, ask for the tests she needed, and keep fighting until the healthcare system responded. “It really saved my life.”

Introducing GPT‑5.4—our most capable and efficient frontier model for professional work. GPT‑5.4 brings together the best of our recent advances in reasoning, coding, and agentic workflows into a single frontier model. It's our most factual and efficient model: fewer tokens, faster speed. GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT. GPT-5.4 is also now available in the API and Codex. https://lnkd.in/gWvcSwZV

GPT-5.3 Instant in ChatGPT is now rolling out to everyone. More accurate, less cringe, and better web search. https://lnkd.in/epW9visG

We’re partnering with Amazon to accelerate AI innovation for enterprises, startups, and end consumers around the world. https://lnkd.in/g7y85Y-9

AI demand is growing rapidly as more people and businesses use AI to learn, build, and get work done every day. Today we’re announcing $110B in new investment, with support from SoftBank Group Corp., NVIDIA, and Amazon, to help scale the infrastructure needed to meet that demand and make AI more accessible, reliable, and useful for everyone. Hundreds of millions of people are using ChatGPT each week, and as usage grows, continued investment in compute, distribution, and partnerships is essential to bringing frontier AI to more communities worldwide. https://lnkd.in/gZ65Cjrh