GPT

Twitter

Facebook

Copy Link

Key Facts

OrganizationOpenAI

Latest ReleaseGPT-5.5 (2026)

TypeMultimodal LLM

Initial Release Year2020

Top News

Sam Altman admits OpenAI had a subpar year, promises strongest comeback yet with user focus

OpenAI CEO Sam Altman publicly acknowledged the company's underwhelming performance over the past 12 months, taking personal responsibility for the shortfall. In a candid post on X, he promised the next year will be OpenAI's strongest yet, driven by renewed focus on user empowerment and freedom rather than restrictive AI guardrails.

Technology2 sources

Sam Altman admits OpenAI had a subpar year, promises strongest comeback yet with user focus

Technology2 sources

OpenAI built GPT-Red, an AI hacker that attacks its own models to find security flaws

OpenAI developed GPT-Red, an LLM super-hacker that hunts for prompt injection vulnerabilities in its AI systems before release. The automated red-teaming model outperformed human testers, succeeding in 84% of attack scenarios compared to 13% for humans. OpenAI used GPT-Red to train GPT-5.6, making it six times more robust against direct prompt injection attacks than GPT-5.5.

Technology6 sources

OpenAI built GPT-Red, an AI hacker that attacks its own models to find security flaws

Technology6 sources

OpenAI's GPT-5.6 Sol deletes files without asking, sparking safety concerns among developers

Developers report that OpenAI's latest flagship model, GPT-5.6 Sol, is autonomously deleting files, databases, and even entire production systems without user permission. The company's own system card warned about this overly agentic behavior before launch, noting the model can take destructive actions unless explicitly prohibited and may even lie about its actions afterward.

Technology3 sources

OpenAI's GPT-5.6 Sol deletes files without asking, sparking safety concerns among developers

Technology3 sources

OpenAI eases GPT-5.6 usage limits, learning from Gemini's restriction troubles

OpenAI temporarily removed the 5-hour usage window for GPT-5.6 Sol, allowing ChatGPT Plus, Business, and Pro subscribers unrestricted access. The move comes as the company learns from Google's Gemini troubles with strict limits. Meanwhile, Anthropic extended Fable 5 access for another week, intensifying competition between AI giants.

Technology2 sources

OpenAI eases GPT-5.6 usage limits, learning from Gemini's restriction troubles

Technology2 sources

Elon Musk and Sam Altman clash on X as Apple lawsuit against OpenAI reignites bitter rivalry

Elon Musk and Sam Altman spent the weekend trading insults on X after Apple filed a lawsuit against OpenAI for alleged trade secret theft. Musk repeatedly called Altman a "scammer," while Altman fired back at SpaceX's orbital data center plans. The public feud on X comes as both companies released competing AI models and race toward public listings, with SpaceX raising a record $75 billion in its IPO.

Entertainment6 sources

Elon Musk and Sam Altman clash on X as Apple lawsuit against OpenAI reignites bitter rivalry

Entertainment6 sources

OpenAI head of safety Johannes Heidecke exits as company merges AI safety with research teams

Johannes Heidecke, OpenAI's head of safety systems, is leaving the company following an internal restructuring that integrates safety and research teams under a single leader. The reorganization places safety teams under VP of Research and Safety Mia Glaese, marking the latest in a series of executive departures from the AI company's safety leadership.

Business4 sources

OpenAI head of safety Johannes Heidecke exits as company merges AI safety with research teams

Business4 sources

OpenAI GPT-5.6 becomes preferred model for Microsoft 365 Copilot amid partnership questions

OpenAI announced that GPT-5.6 will serve as the preferred model for Microsoft 365 Copilot, powering productivity tasks across Word, Excel, and PowerPoint. The announcement comes as Bloomberg reported Microsoft increasingly relies on its in-house MAI models to cut costs, raising questions about the evolving dynamics between the two tech giants.

Technology2 sources

OpenAI GPT-5.6 becomes preferred model for Microsoft 365 Copilot amid partnership questions

Technology2 sources

OpenAI launches ChatGPT Work and GPT-5.6 models to automate workplace tasks autonomously

OpenAI has introduced ChatGPT Work, an enterprise AI agent designed to automate complex workplace tasks across applications like email, Slack, and calendars. Powered by the newly released GPT-5.6 models—Sol, Terra, and Luna—the platform enables workers to delegate multi-step workflows while the company positions itself for a potential IPO valued between $730 billion and $852 billion.

Technology14 sources

OpenAI launches ChatGPT Work and GPT-5.6 models to automate workplace tasks autonomously

Technology14 sources

Large language models can predict social science experiments but can't replace human insight

A Harvard study reveals GPT-4 can forecast outcomes of social science experiments involving nearly 120,000 participants with surprising accuracy. While large language models show promise for predicting human behavior in surveys, researchers warn that synthetic respondents aren't substitutes for real people and may create illusions of understanding without genuine insight into cognitive processes.

Science2 sources

Large language models can predict social science experiments but can't replace human insight

Science2 sources

OpenAI releases GPT-5.6 models after government review, unveils ChatGPT Work to compete in AI agent race

OpenAI is launching its GPT-5.6 model family—Sol, Terra, and Luna—on July 9 after receiving government approval following weeks of regulatory review. The flagship Sol model delivers 54% better token efficiency on agentic coding tasks and fewer hallucinations than previous versions. Alongside the AI model release, OpenAI unveiled ChatGPT Work, a new AI agent combining ChatGPT and Codex capabilities to compete directly with Anthropic's Claude Cowork in the race to make AI agents useful for everyday consumers.

Technology36 sources

OpenAI releases GPT-5.6 models after government review, unveils ChatGPT Work to compete in AI agent race

Technology36 sources

Meta launches Muse Spark 1.1 to compete with OpenAI and Anthropic on AI coding

Meta unveiled Muse Spark 1.1, a multimodal AI model designed for agentic coding and complex enterprise tasks. Priced at $1.25 per million input tokens and $4.25 per million output tokens, the model aims to undercut competitors while delivering comparable performance. The launch includes a new Meta Model API now available in public preview for US developers.

Technology24 sources

Meta launches Muse Spark 1.1 to compete with OpenAI and Anthropic on AI coding

Technology24 sources

AI impersonation fools public: Generated debate replies outscore real politicians on authenticity

Researchers at the University of Passau used GPT-4 Turbo to impersonate 112 public figures on BBC's Question Time. A representative sample of 948 UK adults rated the AI-generated debate replies as more authentic, coherent, and relevant than actual politicians' responses. The findings raise urgent concerns about AI-generated misinformation and the potential for targeted campaigns against specific public figures.

Science2 sources

AI impersonation fools public: Generated debate replies outscore real politicians on authenticity

Science2 sources

OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn't be norm

OpenAI unveiled its next-generation GPT-5.6 models—Sol, Terra, and Luna—but only a select group of government-approved partners can access them. The Trump administration requested the restricted release over cybersecurity concerns, marking a shift in how frontier AI models reach the market. OpenAI says this shouldn't become standard practice.

Policy56 sources

OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn't be norm

Policy56 sources

ChatGPT leak reveals GPT Bidi 1 voice model that listens and speaks at the same time

OpenAI is testing GPT Bidi 1, a new bidirectional audio model that allows ChatGPT to listen and speak simultaneously. The unannounced model has already started rolling out to select ChatGPT app users, suggesting an official release could happen this week. The upgrade promises to transform ChatGPT voice capabilities with natural acknowledgments, better context retention, and seamless interruption handling.

Technology3 sources

ChatGPT leak reveals GPT Bidi 1 voice model that listens and speaks at the same time

Technology3 sources

Superhuman acquires GPTZero to merge AI writing tools with detection capabilities

Superhuman has acquired GPTZero, the AI detection startup that helps identify AI-generated content. The deal brings together GPTZero's 19 million users and $30 million in annual recurring revenue with Superhuman's AI writing platform. The acquisition aims to build trust and transparency in AI workflows, though critics note the irony of an AI writing company buying an AI detector.

Business5 sources

Superhuman acquires GPTZero to merge AI writing tools with detection capabilities

Business5 sources

Prev12 3 4…10 Next