OpenAI's o1 model just outperformed two attending physicians in diagnosing 76 real emergency room cases, according to a new Harvard study published in Science. Tested directly on raw electronic medical records, o1 nailed the diagnosis 67% of the time during triage, topping the doctors' 55% and 50% marks. While critics noted the test used internists instead of ER specialists, researchers say prospective trials in real care settings are urgently needed.
Thinking Machines Lab co-founders Barret Zoph and Luke Metz just left Mira Murati's fledgling startup to rejoin OpenAI, alongside early staffer Sam Schoenholz. The sudden move comes amid allegations that Zoph shared confidential data with competitors, though OpenAI dismissed the concerns and put him in a leadership role. Losing key technical founders is a massive blow to Murati's $12B venture in the middle of a brutal talent war.
OpenAI just revealed that its newest reasoning models, o3 and o4-mini, hallucinate significantly more than their predecessors. On OpenAI's PersonQA benchmark, o3 hallucinated on 33% of questions—about double o1's rate—while o4-mini reached 48%. OpenAI admitted it doesn't fully understand why, noting that scaling reasoning models seems to generate more assertions, both true and false.
OpenAI just rolled out direct PDF inputs for its Responses and Chat Completions APIs. Developers can now pass raw PDF documents straight to models like o1, gpt-4o, and gpt-4o-mini without third-party parsing tools. Why it matters: document workflows just got a lot simpler.
Microsoft AI head Mustafa Suleyman is pushing to reduce the company's heavy dependence on OpenAI, actively testing in-house models like Phi-4 inside Copilot. Microsoft wants to trim sky-high Azure compute bills and reclaim its strategic roadmap, but untangling a deeply integrated $13 billion partnership is proving far easier said than done.
Microsoft just removed all usage limits on Copilot's Voice mode and OpenAI o1-powered Think Deeper reasoning tool for free users. Anyone can now hold uninterrupted voice conversations or work through complex logic problems without hitting query caps. Copilot Pro remains at $20 a month for peak-time priority.
OpenAI just updated the chain of thought for o3-mini, revealing far more reasoning steps to both free and paid ChatGPT users. While it still holds back raw internal monologues to prevent distillation, a new post-processing layer cleans up unsafe output and translates the thought process into non-English native languages.
AI researchers at Stanford and the University of Washington just created an open-source reasoning model called s1 for under $50 in cloud compute. By distilling thinking steps from Google's Gemini into a Qwen base model using 1,000 curated questions, s1 matches frontier reasoning models like OpenAI's o1 on math and coding tests.
OpenAI CEO Sam Altman just admitted that the company is currently losing money on its $200-per-month ChatGPT Pro plan because heavy users are running it far more than anticipated. Altman confessed he personally chose the price thinking it would turn a profit, and OpenAI is now reportedly weighing subscription price hikes or usage-based billing.
Alibaba's Qwen team just released QwQ-32B-Preview, a 32.5B reasoning model licensed under Apache 2.0. Alibaba says it outperforms OpenAI's o1-preview on math benchmarks like AIME and MATH by using inference-time compute, though it can still get caught in loops or stumble on common-sense reasoning.
OpenAI just filed paperwork with the USPTO to trademark "OpenAI o1," moving to shield its new class of reasoning models. Filings reveal the company quietly submitted an application in Jamaica back in May, months before o1 launched. Having previously lost the bid to trademark the generic term "GPT," OpenAI is locking down this brand early.
OpenAI CEO Sam Altman just bet on 2025 as the year we reach artificial general intelligence. In a YC interview, he claimed the path to AGI is now basically clear and boils down to engineering rather than scientific breakthroughs, even as reports swirl about LLM scaling challenges.
OpenAI is reportedly planning to launch Orion, its next frontier AI model, by December, according to The Verge. Sources say it won't land in ChatGPT right away and will go to close partners first, though Sam Altman called the report fake news and an OpenAI spokesperson denied any plans to release Orion this year.
T-Mobile just teamed up with OpenAI to build IntentCX, a customer service platform powered by OpenAI's APIs and the new o1 model. Slated for rollout next year, the system parses support call transcripts to pinpoint customer pain points, aiming well beyond conventional chatbots.
OpenAI just turned its Safety and Security Committee into an independent board oversight group chaired by CMU's Zico Kolter. Crucially, the independent panel now holds the formal authority to delay major model releases until safety concerns are addressed. It marks a significant structural shift to install an external brake on model deployments.
