The Hugging Face Official MCP Server offers unique customization options for AI Assistants accessing the Hub, along with access to thousands of AI applications through one simple URL.
Hugging Face releases ScreenEnv, a Python library for creating and managing isolated Ubuntu desktop environments in Docker containers, designed for testing and deploying GUI/Computer Use agents, with…
Asynchronous inference for robot action prediction and execution decouples these processes to reduce latency and improve responsiveness, achieving a ~2x speed-up in task completion with comparable…
Hugging Face announces that Gradio apps now support the MCP protocol, turning Hugging Face Spaces into an MCP App Store where users can connect specialized AI abilities to any MCP-compatible LLM…
Hugging Face and AMD collaborated to release open-source custom kernels optimized for the AMD MI300X GPU, achieving significant speedups for serving Llama 3.1 405B in FP8 on VLLM.
Hugging Face and Pollen Robotics announced Reachy Mini, an open-source robot for AI experimentation, priced from $399, available in lite and wireless versions, with a Python SDK, integration with…
This blog post describes a five-stage approach to building an efficient multimodal data pipeline, addressing GPU idle time and padding waste during training.
Hugging Face's infrastructure team shares a technical blog post detailing three key alerts used in their production monitoring and alerting system: High NAT Gateway Throughput, Hub Request Logs…
Hugging FaceStory of the weekOpen-weight release· 8 Jul
Hugging Face releases SmolLM3, a fully open 3B small language model that outperforms Llama-3.2-3B and Qwen2.5-3B while being competitive with larger 4B models.
OpenAI joins the EU Code of Practice, committing to responsible AI, and partners with European governments to foster innovation, infrastructure, and economic growth.
OpenAI partners with the American Federation of Teachers to launch a 5-year initiative equipping 400,000 K-12 educators to lead AI innovation in classrooms.
Moonshot AI announces Kimi K2, a Mixture-of-Experts model with 32 billion activated parameters and 1 trillion total parameters, optimized for agentic tasks and open-sourced as Kimi-K2-Base and…
Replicate published a comparison blog post helping users choose the best AI video model by listing specs and features of over 20 models from various providers.
Together AI launched speech-to-text APIs offering high-performance transcription and translation using Whisper V3 Large, claiming 15x faster inference than OpenAI, support for files exceeding 1GB,…
xAI announces Grok 4, described as the most intelligent model with native tool use and real-time search integration. It is available to SuperGrok and Premium+ subscribers and through the xAI API.