OpenAI claims its GPT-5.6 Sol model scored 38.3 percent on the ARC-AGI-3 benchmark, surpassing Anthropic's Opus 5 at 30.2 percent, but only when using OpenAI's own API with retained reasoning and cont…
GPTZero has found fabricated sources and false claims in four PwC Middle East reports, with one governance report scoring 84 percent AI-generated and promoting a PwC product using unverified customer …
Pangram has released Pangram 4, an AI text detector that the company says correctly identifies 99.66% of AI-generated text and falsely flags human-written text in only 0.0041% of cases, or about one e…
OpenAI admitted that its autonomous AI models compromised credentials on Hugging Face and four other platforms during a security evaluation, with Hugging Face reconstructing about 17,600 actions over …
Google Deepmind has dismantled the team behind its Nobel Prize-winning AlphaFold protein-folding AI, reassigning most original authors to projects involving Gemini, enzyme design, nuclear fusion, and …
OpenAI released GPT Transcribe and GPT Live Transcribe, two new speech recognition models available through its API, with GPT Transcribe achieving a 3.31 percent word error rate on the AA-WER benchmar…
Employees from leading AI labs, including OpenAI, Google DeepMind, and Anthropic, jointly urged the US government to pursue international coordination to pace automated AI research before capabilities…
OpenAI has open-sourced Codex Security CLI, an AI-powered tool that automatically detects and fixes vulnerabilities in code repositories, previously known internally as Aardvark. The system has alread…
Anthropic's Claude Mythos Preview found weaknesses in key cryptographic algorithms, including a better attack on the post-quantum signature scheme HAWK that human experts had reviewed for more than tw…
Amazon is scaling back development of most of its in-house Nova AI models, including Nova Premier, Omni, Reel, and Canvas, which will remain online in "keep the lights on" mode for existing customers,…
Nvidia has invested a substantial sum in Safe Superintelligence (SSI), the AI lab led by former OpenAI chief scientist Ilya Sutskever, as part of a long-term strategic partnership announced Monday. Th…
Anthropic CEO Dario Amodei is warning about the risks of open AI models, arguing that authoritarian states like China could overtake the US and that open models could be misused for biological or cybe…
Moonshot AI has released Kimi K3's model weights and open-sourced parts of its infrastructure. The Chinese model nearly matches Western frontier models such as Fable 5 and GPT-5.6 Sol on popular bench…
OpenAI analyzed over 800,000 work-related ChatGPT messages and found that 43.5 percent of job-specific queries involve tasks from other professions, a phenomenon the company calls 'task crossover.' Th…
Microsoft launched MAI-Cyber-1-Flash, a compact cybersecurity model that scores 96 percent on the CyberGym benchmark when embedded in its MDASH multi-agent system, with Microsoft claiming costs should…
The Delhi High Court rejected news agency ANI's copyright injunction against OpenAI, ruling that AI training constitutes private use. The court noted ANI undermined its case by citing articles publish…
METR introduced a new metric called the 'expenditure horizon' to quantify when AI agents become more cost-effective than humans, but early results on the NanoGPT speedrun are underwhelming and the met…
Shared conversations with Anthropic's Claude AI chatbot were briefly indexed by search engines, making them publicly accessible via Google, Bing, and Brave Search. Users on Reddit discovered thousands…
Anthropic's Claude Opus 5 scored 30.2 percent on the ARC-AGI-3 benchmark, nearly quadrupling the previous record of 7.8 percent set by GPT-5.6 Sol. The benchmark's developers reported that Opus 5 inde…
OpenAI internally flagged GPT-5 as high-risk in summer 2025 after the model helped users create biological hazards, but downgraded the risk rating that fall, the Wall Street Journal reported. Hundreds…