✷ The archive · 11 labs · 3,059 dispatches
The labs
No more opening a dozen official blogs every morning. First-hand releases from OpenAI, Anthropic, DeepMind and the rest, each with its substance pulled out.
OpenAI GPT-5 System Card
OpenAI has released GPT-5, a unified system featuring a real-time router that dynamically switches between fast, high-throughput models and deeper reasoning models based on query complexity.
GPT-5 Safe Completions: Transitioning from Refusal-Based to Output-Centric Safety Training
OpenAI has introduced safe-completions for GPT-5, a safety-training approach that maximizes helpfulness while penalizing unsafe outputs, reducing the binary comply-or-refuse trade-off for dual-use prompts.
OpenAI GPT-5 Release Notes
OpenAI has released GPT-5, a unified AI system featuring a real-time router that switches between a fast model and a deeper reasoning model to provide expert-level intelligence across coding, math, and health.
OpenAI and GSA Partnership: ChatGPT Enterprise for U.S. Federal Workforce
OpenAI is providing ChatGPT Enterprise to the entire U.S. federal executive branch workforce for a nominal fee of $1 per agency for one year to reduce administrative burden and improve public service delivery.
Anthropic Appoints Hidetoshi Tojo as Head of Japan and Announces Tokyo Office
Anthropic has appointed Hidetoshi Tojo as Head of Japan and will open its first Asia office in Tokyo to expand local operations and support Japanese enterprises in adopting responsible AI.
OpenAI Estimating Worst Case Frontier Risks of Open Weight LLMs
OpenAI evaluated the risks of releasing gpt-oss by using malicious fine-tuning (MFT) to attempt to elicit maximum capabilities in biology and cybersecurity, finding that the model did not substantially advance the frontier of risk compared to existing open-weight models.
OpenAI gpt-oss-120b and gpt-oss-20b Release
OpenAI has released gpt-oss-120b and gpt-oss-20b, two open-weight reasoning models under the Apache 2.0 license designed for agentic workflows and customizable reasoning effort.
OpenAI gpt-oss Release Notes
OpenAI has released gpt-oss-120b and gpt-oss-20b, two open-weight reasoning models licensed under Apache 2.0 that deliver frontier-level reasoning and tool-use capabilities optimized for consumer hardware.
OpenAI Open Weights Release
OpenAI has released its most capable open-weights reasoning models to democratize AI access and promote US-led democratic AI standards globally.
Anthropic Makes Claude Available via GSA Schedule for U.S. Federal Agencies
Anthropic has made Claude available for purchase through the General Services Administration (GSA) schedule, streamlining procurement for U.S. federal government departments and agencies.
OpenAI gpt-oss Release on Ollama
Ollama has partnered with OpenAI to integrate the gpt-oss open weight models, offering 20B and 120B parameter versions optimized for local execution via the MXFP4 quantization format.
Claude Opus 4.1 Release Notes
Anthropic has released Claude Opus 4.1, an upgrade to Claude Opus 4 that improves performance in agentic tasks, real-world coding, and reasoning.
NVIDIA AI-Q Blueprint: Top-Ranking Open Deep Research Agent on DeepResearch Bench
NVIDIA's AI-Q Blueprint achieves the top spot for open-licensed stacks on the Hugging Face DeepResearch Bench, utilizing a combination of Llama 3.3-70B Instruct and Llama-3.3-Nemotron-Super-49B-v1.5.
Qwen-Image Release: Native Text Rendering and Precise Image Editing
Qwen-Image is a 20B MMDiT image foundation model that provides state-of-the-art complex text rendering and precise image editing capabilities.
OpenAI Optimizing ChatGPT for User Utility and Wellbeing
OpenAI is shifting ChatGPT's optimization goals from engagement metrics like time spent to real-world utility and user wellbeing, introducing break reminders and improved handling of high-stakes personal decisions.
Anthropic Framework for Safe and Trustworthy AI Agents
Anthropic has introduced an early framework for responsible agent development centered on human control, transparency, value alignment, privacy, and security to ensure autonomous AI agents remain safe and reliable.
3LM: A Benchmark for Arabic LLMs in STEM and Code
Hugging Face and TII UAE introduce 3LM, the first comprehensive benchmark designed to evaluate Arabic Large Language Models on STEM subjects and code generation.
Fine-Tuning Pixtral-12B for Satellite Imagery Classification
Mistral AI demonstrates that fine-tuning Pixtral-12B using LoRA on the Aerial Image Dataset (AID) increases classification accuracy from 0.56 to 0.91 while reducing hallucinations from 5% to 0.1%.
Figma AI Integration and Product Evolution
Figma is integrating AI across its platform to automate routine design tasks and introduce prompt-to-app capabilities via Figma Make, shifting the designer's role from implementation to high-level problem solving.
Anthropic Persona Vectors for Monitoring and Controlling AI Character Traits
Anthropic has introduced persona vectors, neural network activity patterns that allow developers to monitor, prevent, and mitigate undesirable character traits like sycophancy and hallucination in language models.
Implementing MCP Servers in Python with Gradio
Hugging Face introduces a method for Python developers to quickly build Model Context Protocol (MCP) servers using Gradio, enabling LLMs to integrate with thousands of AI models and Spaces on the Hugging Face Hub.
OpenAI Introduces Stargate Norway AI Data Center
OpenAI has launched Stargate Norway, a renewable-powered AI data center initiative in Narvik, Norway, aiming to deliver 100,000 NVIDIA GPUs by the end of 2026.
Mistral AI Codestral 25.08 and Enterprise Coding Stack Release
Mistral AI has released Codestral 25.08 and a comprehensive enterprise coding stack featuring integrated completion, semantic search, and agentic workflows with flexible deployment options.
Intercom's Strategy for Sustainable AI Advantage
Intercom achieved a sustainable AI advantage by prioritizing early model fluency, rigorous evaluation frameworks, and a modular architecture that allows for rapid model swapping and cost optimization.
Anthropic Joins CMS Health Tech Ecosystem Pledge for Healthcare Interoperability
Anthropic has signed the Centers for Medicare & Medicaid Services (CMS) Health Tech Ecosystem pledge to use AI to modernize healthcare data sharing and eliminate data silos.
Ollama App Release for macOS and Windows
Ollama has released a new application for macOS and Windows that provides a graphical user interface for downloading and chatting with local AI models, including support for file uploads and multimodal capabilities.
ChatGPT Study Mode Release
OpenAI has introduced study mode in ChatGPT, a learning experience designed to guide students through problems using Socratic questioning and scaffolded responses rather than providing direct answers.
Hugging Face Trackio Release
Hugging Face has released Trackio, a free, open-source, lightweight experiment tracking library that serves as a drop-in replacement for wandb with native integration for Hugging Face Spaces and Datasets.
Qwen GSPO: Scalable Reinforcement Learning for Language Models
Qwen introduces Group Sequence Policy Optimization (GSPO), a sequence-level RL algorithm that improves training stability and efficiency over GRPO, particularly for Mixture-of-Experts (MoE) models.
Hugging Face CLI Update: Transition to hf Command
Hugging Face has renamed the huggingface-cli to hf, introducing a more ergonomic resource-action command structure and a new hf jobs service for running scripts on HF infrastructure.
Parquet Content-Defined Chunking for Efficient Data Deduplication
Hugging Face introduces Parquet Content-Defined Chunking (CDC), now available in PyArrow and Pandas, to enable efficient deduplication of Parquet files on the Xet storage layer, significantly reducing upload and download times.
Qwen-MT Turbo Release Notes
Qwen has released Qwen-MT (qwen-mt-turbo), a lightweight MoE-based translation model supporting 92 languages with high customizability and low API costs.
Outtake Cybersecurity Agents powered by OpenAI
Outtake uses GPT-4.1 and OpenAI o3 to automate the detection and resolution of digital threats, reducing takedown timelines from 60 days to a few hours.
Fast LoRA Inference for Flux with Diffusers and PEFT
Hugging Face introduces an optimization recipe for Flux.1-Dev that achieves up to 2.23x speedup in LoRA inference by combining hotswapping, torch.compile, and FP8 quantization.
TimeScope: A New Benchmark for Long-Video Large Multimodal Model Understanding
Hugging Face introduces TimeScope, an open-source benchmark that evaluates the temporal comprehension of vision-language models using video needles inserted into content ranging from 1 minute to 8 hours.
OpenAI DevDay 2025 Announcement
OpenAI will host its third annual DevDay on October 6, 2025, in San Francisco to share new developments and connect over 1,500 developers.
Model ML AI Infrastructure for Financial Services
Model ML is leveraging OpenAI's latest models and Agent SDK to provide financial firms with purpose-built AI agents and end-to-end workflow automation.
Anthropic Response to America's AI Action Plan
Anthropic supports the White House's AI Action Plan for its focus on infrastructure and federal adoption but urges stronger national transparency standards and stricter export controls on H20 chips to China.
Anthropic and University of Chicago BFI Partnership for AI Economic Research
Anthropic has partnered with the University of Chicago's Becker Friedman Institute for Economics (BFI) to study AI's impact on labor markets and the economy using Claude for Enterprise and the Anthropic Economic Index.
Mistral AI Environmental Impact Study and Proposed Global Standards
Mistral AI has conducted a first-of-its-kind comprehensive lifecycle analysis of its LLMs, disclosing specific carbon, water, and resource depletion metrics for Mistral Large 2 to advocate for a global environmental reporting standard.
Qwen3-Coder Release Notes
Qwen has released Qwen3-Coder, a Mixture-of-Experts model that achieves state-of-the-art open-model performance in agentic coding, browser-use, and tool-use, comparable to Claude Sonnet 4.
OpenAI and Penda Health AI Clinical Copilot Study
OpenAI and Penda Health developed AI Consult, an LLM-powered copilot that reduced diagnostic errors by 16% and treatment errors by 13% across nearly 40,000 patient visits in Nairobi, Kenya.
OpenAI and Oracle Partner to Expand Stargate AI Infrastructure by 4.5 GW
OpenAI and Oracle have partnered to develop 4.5 gigawatts of additional Stargate data center capacity in the U.S., bringing OpenAI's total under-development capacity to over 5 GW.
OpenAI Economic Analysis and Productivity Research Initiative
OpenAI has released its first economic analysis on ChatGPT's productivity gains and launched a 12-month research collaboration to assess AI's impact on the workforce.
NVIDIA NIM Integration with Hugging Face
NVIDIA has expanded NVIDIA NIM to support over 100,000 LLMs on Hugging Face, providing a single Docker container that automates model analysis, backend selection, and performance optimization for rapid deployment.
OpenAI and UK Government Strategic Partnership for AI-Driven Growth
OpenAI and the UK Government have signed a Memorandum of Understanding to accelerate AI adoption across public and private sectors and expand AI infrastructure in the UK.
OpenAI CEO of Applications Announcement: AI as a Tool for Global Empowerment
OpenAI has appointed a new CEO of Applications to lead the effort in democratizing access to AI-driven knowledge, healthcare, economic freedom, and personal support.
Anthropic to sign the EU General-Purpose AI Code of Practice
Anthropic has announced its intention to sign the EU General-Purpose AI Code of Practice to advance transparency, safety, and accountability in frontier AI development.
Anthropic Build AI in America Energy Report
Anthropic proposes a strategic framework to secure 50GW of electric capacity by 2028 to maintain U.S. AI leadership and compete with China's rapid energy infrastructure expansion.
OpenAI Launches $50 Million Fund for Nonprofit and Community Organizations
OpenAI has established an initial $50 million fund to provide direct support to nonprofit and community organizations to leverage AI for public good in sectors like healthcare, education, and economic opportunity.