The Archive

Anthropic publishes compliance framework for California's Transparency in Frontier AI Act regulatory requirements.

Anthropic partners with US Department of Energy to apply AI to scientific research and discovery.

Anthropic outlines user safety and wellbeing protections for Claude platform.

OpenAI introduces chain-of-thought monitorability framework across 13 evaluations showing internal reasoning monitoring outperforms output monitoring.

OpenAI updates Model Spec with Under-18 Principles for age-appropriate teen guidance grounded in developmental science.

OpenAI and U.S. Department of Energy sign MOU to collaborate on AI and advanced computing for scientific discovery.

OpenAI releases AI literacy resources for teens and parents covering responsible use, critical thinking, and healthy boundaries.

University of Pennsylvania applies DINO and SAM to automate medical triage and emergency response workflows.

Meta's Segment Anything Model deployed by Universities Space Research Association for flood emergency response via USGS water monitoring systems.

GPT-5.2-Codex is OpenAI's most advanced coding model with long-horizon reasoning, large-scale code transformations, and cybersecurity capabilities.

OpenAI releases system card addendum for GPT-5.2-Codex, detailing safety and capability documentation for advanced coding model.

OpenAI releases GPT-5.2-Codex with long-horizon reasoning, large-scale code transformations, and cybersecurity capabilities.

Mistral releases OCR 3, advancing accuracy and efficiency benchmarks for document processing and layout analysis.

Google DeepMind releases Gemini 3 Flash, a frontier-grade model optimized for speed and cost efficiency.

OpenAI launches OpenAI Academy for News Organizations with American Journalism Project and Lenfest Institute to train newsrooms on responsible AI adoption.

Grok Voice Agent API available to developers, extending voice capabilities.

OpenAI publishes data-driven analysis of enterprise AI adoption patterns, tracking progression from experimentation to productivity gains.

OpenAI enables developer submissions for ChatGPT app directory with updated SDK, guidelines, and in-product discovery.

Open interpretability tools for language models are now available across the entire Gemma 3 family with the release of Gemma Scope 2.

OpenAI introduces FrontierScience benchmark measuring AI reasoning across physics, chemistry, and biology research tasks.

OpenAI presents evaluation framework using GPT-5 to measure AI acceleration of wet-lab biological research, including molecular cloning optimization.

Meta introduces SAM Audio, unified multimodal model for audio source separation accepting text, visual, and temporal prompts.

OpenAI publishes organizational guidance on AI readiness covering strategy, training, governance, and innovation acceleration.

OpenAI upgrades ChatGPT Images with GPT-Image-1.5, delivering 4× faster generation, precise edits, and consistent detail.

BBVA deploys ChatGPT Enterprise across 120,000 employees via multi-year partnership with OpenAI for banking transformation.

OpenAI demonstrates rapid Android development: Sora shipped in 28 days using Codex for AI-assisted planning and parallel coding workflows.