📰 Hugging Face Blog

Introducing NVIDIA Nemotron 3 Nano Omni: Long-Context Multimodal Intelligence for Documents, Audio and Video Agents

**Introducing SPEED-Bench: A Unified and Diverse Benchmark for Speculative Decoding**

Holotron-12B - High Throughput Computer Use Agent

Bringing Robotics AI to Embedded Platforms: Dataset Recording, VLA Fine‑Tuning, and On‑Device Optimizations

Introducing Modular Diffusers - Composable Building Blocks for Diffusion Pipelines

Train AI models with Unsloth and Hugging Face Jobs for FREE

IBM and UC Berkeley Diagnose Why Enterprise Agents Fail Using IT-Bench and MAST

Custom Kernels for All from Codex and Claude

We Got Claude to Build CUDA Kernels and teach open models!

Unlocking Agentic RL Training for GPT-OSS: A Practical Retrospective

AprielGuard: A Guardrail for Safety and Adversarial Robustness in Modern LLM Systems

Introducing swift-huggingface: The Complete Swift Client for Hugging Face

DeepMath: A lightweight math reasoning Agent with smolagents

Transformers v5: Simple model definitions powering the AI ecosystem

Diffusers welcomes FLUX-2

Continuous batching from first principles

Building Deep Research: How we Achieved State of the Art

20x Faster TRL Fine-tuning with RapidFire AI

Voice Cloning with Consent

LeRobot v0.4.0: Supercharging OSS Robot Learning

SOTA OCR with Core ML and dots.ocr

Accelerating Qwen3-8B Agent on Intel® Core™ Ultra with Depth-Pruned Draft Models

VibeGame: Exploring Vibe Coding Games

Swift Transformers Reaches 1.0 – and Looks to the Future