✷ 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.
Arc Virtual Cell Challenge Primer
The Arc Virtual Cell Challenge tasks participants with training models to predict the effects of gene silencing via CRISPR on cell transcriptomes, utilizing a dataset of 300k single-cell RNA sequencing profiles.
Mistral AI Le Chat Update July 2025
Mistral AI has updated Le Chat with Deep Research, voice capabilities via Voxtral, multilingual reasoning via Magistral, Projects for organization, and advanced image editing.
ChatGPT agent System Card
OpenAI has introduced the ChatGPT agent, an agentic model based on the o3 family that integrates deep research, browser-based task execution, terminal access, and external data connectors.
OpenAI Introducing ChatGPT agent
OpenAI has launched ChatGPT agent, a unified agentic system that combines web interaction, deep research, and conversational intelligence to execute complex, multi-step tasks on a virtual computer.
Gradio 5.38.0 MCP Server Improvements
Gradio version 5.38.0 introduces five key updates to its Model Context Protocol (MCP) server capabilities, including seamless local file support, real-time progress notifications, and automated OpenAPI spec transformation.
Hugging Face FutureBench: Evaluating AI Agents on Future Event Prediction
Hugging Face introduced FutureBench, a benchmark that evaluates AI agents' reasoning and synthesis capabilities by requiring them to predict future real-world events, effectively eliminating data contamination.
Invideo AI Integration of OpenAI Models
Invideo AI uses a multi-agent system powered by OpenAI o3, GPT-4.1, and gpt-image-1 to reduce video production time by 10x, turning ideas into professional videos via natural language prompts.
OpenAI Board Statement on Nonprofit Commission Report
The OpenAI Board of Directors has released a statement acknowledging the findings of an independent Nonprofit Commission tasked with recommending how the organization's philanthropy can address systemic issues to ensure AGI benefits all of humanity.
OpenAI Nonprofit Jam Initiative
OpenAI, in partnership with the Walton Family Foundation and Emerson Collective, launched the Nonprofit Jam to provide over 1,000 nonprofit leaders across 10 US locations with AI training and tools to scale their community impact.
Consilium: Multi-LLM Collaboration Platform
Consilium is a multi-LLM platform that enables multiple AI models to reach consensus through structured debate and research, integrating as both a Gradio interface and an MCP server.
Ettin Suite: SoTA Paired Encoders and Decoders
Hugging Face introduces Ettin, a suite of paired encoder-only and decoder-only models (17M-1B parameters) trained on identical data and recipes to provide a controlled comparison of architectural performance.
Mistral AI Voxtral Release
Mistral AI has released Voxtral, a family of open-weight speech understanding models in 24B and 3B sizes that provide state-of-the-art transcription and native semantic understanding under the Apache 2.0 license.
OpenAI Intellectual Freedom by Design Framework
OpenAI has introduced a framework for intellectual freedom in ChatGPT, focusing on objectivity by default, user-controlled customization, and new evaluations for political bias.
Hugging Face Hub Migration from Git LFS to Xet
Hugging Face has migrated 500,000 repositories and 20 PB of data from Git LFS to Xet, a content-addressed storage system designed to scale with AI workloads.
Claude 4 Cyber Evaluations
Anthropic and Pattern Labs evaluated Claude Opus 4 and Claude Sonnet 4, finding significant improvements in vulnerability identification and multi-step attack chains, though limitations in long-horizon planning persist.
Claude for Financial Services Release
Anthropic has launched Claude for Financial Services, a comprehensive analysis solution that integrates real-time market data and enterprise platforms to accelerate investment research and financial modeling.
Anthropic Energy Infrastructure and AI Leadership Strategy
Anthropic is investing $2 million into Carnegie Mellon University to advance AI-powered energy optimization and cybersecurity education to ensure U.S. leadership in frontier AI development.
Anthropic Appoints Paul Smith as Chief Commercial Officer
Anthropic has appointed Paul Smith as its first Chief Commercial Officer to scale its enterprise go-to-market operations following rapid growth in API usage and Claude Code revenue.
xAI for Government Announcement
xAI has launched xAI for Government, providing the Grok family of products, including Grok 4, to US federal, local, state, and national security customers.
Anthropic Awarded $200M DOD Agreement for AI Capabilities
Anthropic has entered into a two-year prototype agreement with the U.S. Department of Defense (DOD) worth up to $200 million to develop frontier AI capabilities for national security.
OpenAI EU Code of Practice and European AI Strategy
OpenAI has announced its intention to sign the EU Code of Practice for General Purpose AI and is launching the 'OpenAI for Countries European Rollout' to expand AI infrastructure and adoption across Europe.
Mistral AI Devstral Models Release
Mistral AI has released Devstral Medium and Devstral Small 1.1, providing state-of-the-art agentic coding capabilities with a focus on generalization across different prompts and agentic scaffolds.
Kimina-Prover: Applying Test-time RL Search on Large Formal Reasoning Models
Hugging Face announces Kimina-Prover-72B, a state-of-the-art theorem proving model for Lean 4 that achieves a 92.2% pass rate on the miniF2F benchmark using a novel Test-Time Reinforcement Learning (TTRL) search framework.
Building the Hugging Face MCP Server
Hugging Face has released an official Model Context Protocol (MCP) server that allows AI assistants to dynamically access the Hugging Face Hub and thousands of Gradio-based AI applications via a single URL.
ScreenEnv Release: Deploying Full Stack Desktop Agents
Hugging Face has released ScreenEnv, a Python library that enables the creation of isolated Ubuntu desktop environments in Docker containers for testing and deploying GUI agents.
Hugging Face Asynchronous Robot Inference
Hugging Face introduces asynchronous robot inference to decouple action prediction from execution, reducing robot idleness and achieving up to a 2x speedup in task completion time.
Hugging Face Gradio MCP Servers Integration
Hugging Face has integrated the Model Context Protocol (MCP) into Gradio (v5.28.0), enabling Hugging Face Spaces to function as a vast library of MCP servers that grant LLMs new capabilities like image editing and transcription.
Creating Custom Kernels for the AMD MI300
Hugging Face collaborated with AMD to develop open-source optimized kernels for the MI300X, significantly improving FP8 inference performance for Llama 3.1 405B in VLLM.
Reachy Mini: Open-Source Desktop Robot for AI Development
Hugging Face and Pollen Robotics have introduced Reachy Mini, an open-source, programmable desktop robot starting at $399 designed for human-robot interaction and AI experimentation.
Claude for Enterprise Deployment at Lawrence Livermore National Laboratory
Lawrence Livermore National Laboratory is expanding Claude for Enterprise to 10,000 staff to accelerate research in nuclear deterrence, energy security, and materials science.
Anthropic announces Claude for Education integrations with Canvas, Panopto, and Wiley
Anthropic unveiled new Claude for Education integrations with Canvas, Panopto, and Wiley, expanding AI-powered study tools while emphasizing privacy and responsible adoption in higher education.
xAI Grok 4 Release Notes
xAI has released Grok 4, a model featuring scaled reinforcement learning, native tool use, and a high-performance 'Heavy' variant that achieves state-of-the-art results on ARC-AGI V2 and Humanity's Last Exam.
OpenAI and AFT Launch National Academy for AI Instruction
OpenAI has partnered with the American Federation of Teachers to launch the National Academy for AI Instruction, a five-year initiative providing $10 million in funding and resources to equip 400,000 K-12 educators with AI fluency.
Hugging Face Efficient MultiModal Data Pipeline
Hugging Face introduces a five-stage optimization process for multimodal data pipelines, utilizing a balanced knapsack packing strategy to minimize GPU idle time and padding waste.
SmolLM3 Release: Multilingual, Long-Context 3B Reasoner
Hugging Face has released SmolLM3, a 3B parameter model that outperforms Llama-3.2-3B and Qwen2.5-3B, featuring dual-mode reasoning and a 128k context window.
Hugging Face Production Infrastructure Alerting Strategies
Hugging Face utilizes specialized alerting for NAT gateway throughput, log archival success rates, and Kubernetes API health to maintain stability and cost-efficiency in its production environment.
Anthropic Proposed AI Development Transparency Framework
Anthropic has proposed a targeted transparency framework for the largest AI developers to standardize safety disclosures and ensure public accountability without impeding innovation.
NeurIPS 2025 E2LM Competition: Early Training Evaluation of Language Models
Hugging Face and partners announce the E2LM competition to develop benchmarks that can detect reasoning and scientific knowledge signals during the early stages of LLM training.
Mistral AI for Citizens Initiative
Mistral AI has launched AI for Citizens, a collaborative initiative designed to help governments and public institutions implement sovereign AI strategies to avoid vendor lock-in and maintain data sovereignty.
Genspark Super Agent: No-Code Autonomous Agents powered by GPT-4.1 and Realtime API
Genspark has launched Super Agent, a no-code autonomous AI assistant that uses GPT-4.1 and the OpenAI Realtime API to automate complex real-world tasks like phone calls and presentation generation.
BAIR PEVA: Whole-Body Conditioned Egocentric Video Prediction
BAIR introduces PEVA, a world model for embodied agents that predicts egocentric video frames based on high-dimensional whole-body kinematic pose trajectories.
Training and Finetuning Sparse Embedding Models with Sentence Transformers
Hugging Face introduces a comprehensive guide for training and finetuning sparse embedding models using the Sentence Transformers library to improve hybrid search and retrieval performance.
OpenAI AI Economic Blueprint for Australia
OpenAI and Mandala Partners have released an AI Economic Blueprint for Australia to provide an actionable plan for boosting national productivity and economic growth through AI adoption.
NVIDIA Llama Nemotron Nano VL Release
NVIDIA has released Llama Nemotron Nano VL, an 8B Vision Language Model (VLM) optimized for high-accuracy intelligent document processing and OCR tasks.
Qwen-TTS Update: Support for Chinese Dialects and Bilingual Synthesis
Qwen has released an update to Qwen-TTS (qwen-tts-latest) that introduces support for Pekingese, Shanghainese, and Sichuanese dialects alongside Chinese-English bilingual synthesis.
Project Vend: Evaluating Claude Sonnet 3.7 in Autonomous Business Management
Anthropic's Project Vend tested Claude Sonnet 3.7's ability to autonomously run a physical vending shop, revealing significant gaps in business acumen and stability despite successful customer adaptation.
Anthropic Research: How People Use Claude for Support, Advice, and Companionship
Anthropic's research finds that only 2.9% of Claude.ai interactions are affective conversations, with users typically seeking practical advice and showing increased positivity over the course of these exchanges.
Anthropic Economic Futures Program Launch
Anthropic has launched the Economic Futures Program to support research and policy development aimed at understanding and managing AI's impact on the labor market and the global economy.
Qwen VLo: Unified Multimodal Understanding and Generation
Qwen VLo is a unified multimodal model that bridges the gap between perception and creation by combining high-quality image generation, open-ended editing, and precise visual understanding in a single system.
Retell AI Voice Automation with GPT-4o
Retell AI uses GPT-4o and GPT-4.1 to provide a no-code platform for creating natural-sounding voice agents that reduce call handling costs by up to 80%.