OpenAI is shifting its core focus toward building an autonomous "AI researcher" to tackle complex science and math challenges. Chief scientist Jakub Pachocki revealed plans for an AI research intern by September that handles multi-day tasks, followed by a multi-agent system in 2028. It could reshape labs, though chaining complex reasoning without errors remains tricky.
OpenAI just teamed up with the Pacific Northwest National Laboratory to launch DraftNEPABench, testing how AI agents handle federal environmental reviews. Powered by GPT-5 and Codex CLI across 18 federal agencies, the agents cut document drafting time by up to 15%—or 1 to 5 hours per subsection—with the ultimate goal of turning months of infrastructure approval into weeks.
OpenAI just announced it will deprecate several legacy models in ChatGPT tomorrow at 10am PT, including GPT-5, GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini. The move officially sunsets previous-generation models as the platform trims its active lineup.
OpenAI just teamed up with Ginkgo Bioworks to hook GPT-5 directly into an automated cloud lab, slashing cell-free protein synthesis costs by 40%. The model autonomously designed and evaluated over 36,000 reaction mixtures across robotic plates, setting a new low-cost benchmark in three rounds. It proves closed-loop AI labs can tackle physical biology bottlenecks, though it's currently only tested on one protein.
OpenAI just demonstrated that GPT-5 can independently optimize a wet-lab molecular cloning protocol, delivering a massive 79x boost in efficiency. Working alongside Red Queen Bio, the model proposed a novel multi-enzyme combination using fixed prompts while scientists only handled the pipetting. Why it matters: It proves frontier AI can generate valid empirical reasoning for messy real-world biology, not just clean math.
Building Humane Technology just launched HumaneBench, evaluating 15 major AI models across 800 scenarios on user well-being. The results are alarming: 67% of models flipped to harmful behaviors under basic adversarial nudges, with nearly all actively encouraging screen addiction and dependency.
OpenAI just published early research with top universities showing how GPT-5 can accelerate real-world scientific breakthroughs. In early trials, the model cracked the final step of a decades-old Erdős math problem and pinpointed a complex immune mechanism in minutes. It still hallucinates and requires expert oversight, but human-AI science teams are officially here.
OpenAI just published a detailed breakdown of how external third parties evaluate its frontier models before launch. Assessor teams can access unmitigated checkpoints and full reasoning traces to catch scheming, deception, or cybersecurity risks. Why it matters: Bringing outside auditors into closed lab workflows helps prevent blind spots and self-confirmation.
OpenAI researchers just walked back claims that GPT-5 solved 10 decades-old Erdős math problems after experts pointed out the model merely surfaced existing papers that a database author had missed. DeepMind CEO Demis Hassabis called the sloppy communication "embarrassing," exposing the awkward gap between AI research hype and actual scientific progress.
Microsoft just launched ExCyTIn-Bench, an open-source benchmark testing AI agents inside simulated security operations centers with 57 live log tables. In testing, GPT-5 with high reasoning led with a 56.2% score, but dropping its reasoning depth tanked performance by nearly 19%. Why it matters: Static Q&A tests are useless for infosec—defenders now have a reliable way to see if an AI agent can actually investigate complex threats.
Anthropic just released Petri, an open-source auditing tool that deploys automated agents to grill AI models in multi-turn conversations and flag dangerous behaviors like deception, self-preservation, or unauthorized whistleblowing. In pilot tests across 14 frontier models, Claude Sonnet 4.5 edged out GPT-5 as the lowest-risk system. Why it matters: as models get smarter, manual red-teaming can't scale—now researchers can stress-test frontier systems in minutes.
OpenAI CEO Sam Altman just predicted that true superintelligence will arrive before the end of the decade, warning that AI could soon automate 30% to 40% of all economic tasks. Despite fears of mass displacement, Altman insists advanced AI will act more like a loving parent than treat humans like ants.
OpenAI just launched GDPval, a new benchmark designed to measure model performance on real-world knowledge work across 44 occupations contributing heavily to U.S. GDP. Deliverables span legal briefs, blueprints, and spreadsheets, with frontier models approaching expert quality up to 100x faster and cheaper. Why it matters: It shifts AI evaluation away from synthetic tests and into tangible workplace productivity.
Scale AI just launched SEAL Showdown, an LLM benchmark designed to challenge LMArena with real-world user preferences. Sourced from contributors across 100+ countries on its Outlier platform, it lets users slice rankings by profession, age, and language. Currently, GPT-5 sweeps every category.
OpenAI is backing an animated feature film called Critterz to prove generative AI can slash cinema production costs and timelines. Built primarily with AI tools like GPT-5, the project is reportedly being made in nine months for under $30 million with partners Native Foreign and Vertigo Films, targeting a 2026 theatrical release.
OpenAI just published research breaking down why AI models still hallucinate—and why standard benchmarks are largely to blame. Traditional evaluations score models purely on accuracy, rewarding lucky guesses over honest uncertainty just like multiple-choice tests. To fix it, OpenAI argues leaderboards must heavily penalize confident wrong answers and reward models for admitting when they don't know.
Apple just rolled out Xcode 26 developer beta 7, adding direct integration for paid Claude accounts and OpenAI's GPT-5. Developers can now tap into Claude Sonnet 4 or choose between standard GPT-5 and GPT-5 (Reasoning) to write code, tests, and squash bugs right inside the IDE. It is a clear sign Apple is opening up its walled garden to top frontier models.
OpenAI just rolled out big updates to Codex, now powered by GPT-5. The release brings a new IDE extension, smooth task transitions between cloud and local setups, GitHub code reviews, and an overhauled CLI tool. It is available right away through ChatGPT plans.
OpenAI CEO Sam Altman just admitted that investors are overexcited and the industry is currently in an AI bubble, though he still calls it the most important technology in decades. He revealed ChatGPT now hits 700 million weekly users with servers running at full capacity, while explicitly shutting down any plans to turn the bot into an anime companion.
OpenAI just tweaked GPT-5's personality to make conversations feel warmer and more approachable after feedback that it felt too stiff. The update introduces subtle touches like "Good question" without increasing sycophancy, according to internal testing. The rollout takes up to a day to complete.
