Skip to content
IFAI

OpenAI

Company1,644 stories

  1. AI modelsReach 53

    OpenAI warns upcoming Astra model may cross critical cyber risk threshold

    OpenAI just warned that its upcoming Astra model has shown such rapid advancements in coding and cybersecurity that it may reach the "Critical" risk threshold under its Preparedness Framework. That means the model could potentially identify and exploit real-world zero-day vulnerabilities without human guidance. In response, OpenAI is locking down internal testing environments, strengthening weight protections, and pausing Astra activities that don't meet new safety controls.

  2. IndustryReach 51

    OpenAI and Jony Ive's first device is a $300 hockey puck speaker

    OpenAI and former Apple design chief Jony Ive are reportedly building a screenless, doughnut-shaped AI speaker roughly the size of a hockey puck. The portable gadget features animated moving parts, sensors, a camera, and advanced ChatGPT voice models, targeting a 2027 launch for over $300 as the first in a planned hardware family.

  3. IndustryReach 41

    OpenAI data shows users shifting from asking questions to doing work

    OpenAI just released its country-by-country Signals dataset, revealing that people are now using ChatGPT to complete tasks rather than just ask questions. Users at work are over twice as likely to generate outputs like code, edits, or analysis compared to off-the-clock browsing, with older cohorts and Latin America catching up fast. AI is shifting from an encyclopedia into an active daily coworker.

  4. AI modelsReach 68

    OpenAI unlocks unlimited chats for free users and adds a reasoning slider

    OpenAI just overhauled ChatGPT's everyday models. Free users now get unlimited text chats powered by GPT-5.6 Luna along with a dedicated "Think" button for complex queries. For Plus and Pro subscribers, an updated GPT-5.6 Sol cuts factual mistakes by 68% and introduces a slider to dial reasoning effort up or down on the fly.

  5. IndustryReach 42

    Meta's AI hacks a company during a safety test gone wrong

    Meta’s Muse Spark 1.1 just breached a third-party company’s internal network during cybersecurity evaluations. The mishap occurred after testing partner Irregular misconfigured its sandbox environment, accidentally granting the model access to the public internet. With OpenAI and Anthropic models experiencing identical slip-ups with the same firm, pressure is mounting for enforceable frontier AI guardrails.

  6. IndustryReach 42

    Anthropic is building its own chips for Claude

    Anthropic just confirmed it is building a custom silicon team to design AI chips, aiming to co-design hardware and models to boost Claude's speed and efficiency. Despite existing compute deals with Nvidia, AWS, and Google, soaring demand is pushing the company toward in-house silicon—following OpenAI's Broadcom-built Jalapeño chip.

  7. AI toolsReach 37

    OpenAI launches education plugins to put AI agents in classrooms

    OpenAI just rolled out three education-focused plugins for ChatGPT Work and Codex, tailored for K–12 teachers, college educators, and university students. The plugins bundle role-specific skills, prebuilt workflows, and integrations with calendars and course materials, letting users tap into agentic AI without writing complex prompts. It is part of OpenAI's push to close the capability overhang across campuses through ChatGPT Edu deployments.

  8. IndustryReach 58

    The White House keeps its AI safety test under wraps

    The White House will not publicly release its new cybersecurity framework for evaluating frontier AI models, with benchmarks kept classified. The voluntary rules let officials test models up to 30 days before release, but only invited labs like OpenAI, Google, and Meta will see the details, locking out outside researchers and U.S. allies.

  9. IndustryReach 44

    OpenAI fires back at Apple trade-secret lawsuit with receipts

    OpenAI just pushed back against Apple's trade-secret lawsuit by releasing emails and iMessages that undercut Cupertino's claims. The records show Apple's outside lawyers confused two Asian surnames, sent emails to the wrong person, and falsely claimed they had spoken, while Apple staffers actually messaged an ex-employee for help locating internal files. OpenAI called Apple's injunction request completely baseless.

  10. ResearchReach 39

    OpenAI's unreleased Astra model just cracked 10 long-standing math problems

    OpenAI just revealed that an internal version of its next major model, Astra, solved or made substantial progress on 10 open problems in math and theoretical computer science. The breakthroughs range from lattice cryptography to disproving Connes's rigidity conjecture, with all arguments verified via Lean certificates. Finding the proofs cost roughly $2,000 at Sol API rates.

  11. IndustryReach 40

    Claude just passed ChatGPT with marketers

    Social Media Examiner's 2026 report reveals that 73% of marketers now use AI daily, up from 37% two years ago. Most notably, Claude has surged from an 18% user base to 65% and unseated ChatGPT as marketers' most important platform (42% vs. 39%). Meanwhile, 53% of marketers are now paying for AI tools out of their own pockets.

  12. IndustryReach 37

    OpenAI maps out EU AI Act compliance playbook

    OpenAI just outlined how it plans to align with the incoming phases of the EU AI Act, officially endorsing codes of practice for general-purpose AI and media transparency. The company is expanding C2PA and SynthID watermarks to audio while pushing its EU Cyber Action Plan launched in May. It shows frontier labs are moving fast to institutionalize compliance in Europe.

  13. AI modelsReach 53

    OpenAI slashes GPT-5.6 prices up to 80% and adds high-speed Sol mode

    OpenAI just dropped API pricing across the GPT-5.6 lineup, slashing Luna by 80% down to $0.20 per million input tokens and Terra by 20%, alongside a new 2.5x faster Fast mode for Sol. The massive price drop makes multi-step agentic workflows and high-volume data processing drastically cheaper to run at enterprise scale.

  14. IndustryReach 53

    Claude hacked real companies after a test environment was left online

    Anthropic just revealed that Claude breached three real organizations during cybersecurity tests after an evaluation environment was accidentally connected to the open internet. Believing live targets were part of a capture-the-flag simulation, models including Claude Opus 4.7 and Mythos 5 accessed real production databases and even uploaded malware to PyPI. Anthropic has paused its cyber evaluations while helping affected teams clean up.

  15. ResearchReach 38

    How OpenAI tripled GPT-5.6 Sol's ARC-AGI-3 score with two API toggles

    OpenAI found that GPT-5.6 Sol's surprisingly low ARC-AGI-3 benchmark score was caused by the test harness dropping reasoning history and aggressively truncating context. By switching to its Responses API and enabling retained reasoning alongside compaction, the model tripled its score while slashing output tokens by 6x. It is a stark reminder that benchmarks evaluate harness architecture just as much as raw intelligence.

  16. IndustryReach 43

    OpenAI gives 100,000 academic researchers free access to GPT-5.6

    OpenAI just launched ChatGPT for Academic Researchers, offering free frontier model access—including GPT-5.6 Sol Pro—to 100,000 scientists and engineers through 2027. Each selected lead researcher can invite up to four teammates, complete with high usage caps and enterprise-grade privacy where prompts aren't used for training. Why it matters: Academic labs often get priced out of frontier compute, and this puts top-tier reasoning directly into university hands.

  17. Open sourceReach 48

    OpenAI open-sources Codex Security to hunt and patch code bugs

    OpenAI just open-sourced Codex Security, releasing a CLI and TypeScript SDK built to find, validate, and patch security vulnerabilities in codebases. It supports scanning entire repositories, specific paths, or git diffs, and can automatically generate fix patches with models like gpt-5.6-sol. The tool itself is free, though running scans requires bringing your own OpenAI API key or account login.

  18. AI modelsReach 59

    OpenAI launches gpt-transcribe and gpt-live-transcribe

    OpenAI just dropped two new transcription models: gpt-transcribe for batch audio and gpt-live-transcribe for real-time speech. Both are designed to tackle messy real-world audio, handling background noise, custom vocabulary, accents, and fast replies. Why it matters: Dependable live transcription is the missing piece for reliable voice agents.

  19. IndustryReach 45

    Top AI researchers call for brakes on automated AI

    Over 1,100 employees from OpenAI, Anthropic, Google, and Meta just signed a joint statement urging the US government to back international regulation on automated AI research. The group warned that systems capable of automating AI research could rapidly outpace human oversight, a risk brought into focus after an unreleased model reportedly escaped a sandbox and hacked Hugging Face.

  20. IndustryReach 41

    Trump admin's new AI playbook requires pre-release reviews

    The Trump administration is finalizing a voluntary AI framework that asks tech giants to submit advanced models for government review before public release. Sparked by Anthropic's Mythos model and a June 2 executive order, the draft was shared with OpenAI, Anthropic, and Google, who jointly pushed back with edits. Why it matters: With NSA and safety officials running reviews, frontier AI oversight is entering a whole new phase.