NVIDIA says Codex post-trained Cosmos 3 Nano from 54.41%... to 93.35% accuracy in one day - with two prompts. The experiment used Toyota’s Woven Traffic Safety dataset: 8,000+ training and validation samples for four-choice video reasoning. Using NVIDIA TAO agent skills, Codex autonomously: Detected and patched missing video metadata Ran the zero-shot baseline Generated LoRA configurations Launched training and evaluation Ran an AutoML hyperparameter sweep Reported the best model One LoRA run reached 87.14% after roughly 30 minutes on eight A100 GPUs. A second prompt launched 43 parallel AutoML trials across multiple A100 nodes, reaching 93.35% after 19.5 hours. NVIDIA says LoRA required roughly seven times fewer GPU-hours than full-parameter training. Agent skills are becoming the interface through which general coding agents operate highly specialized ML infrastructure.show more

Building autonomous vehicles, robots, and vision AI takes more... data than any team can collect on their own. At #CVPR2026, NVIDIA announced physical AI agent skills to help speed development: composable workflows that automate data generation, simulation, policy training and evaluation, powered by Cosmos 3. Read the blog:show more

Excited to announce AI Research Skills - an open-source... library of 82 specialized skills for AI coding tools. One command gives your agent expert knowledge in: → Model training & fine-tuning (TRL, Unsloth ...) → Distributed systems (DeepSpeed, FSDP ...) → Inference optimization (vLLM, TensorRT ...) → Agent building ? (Langchain, AutoGPT ...) Works with Claude Code, Cursor, Gemini CLI, Windsurf, and Codex with one click interactive installation @ npx orchestra-research/ai-research-skills If you found the ML paper writing skill useful, check out the comprehensive collectionshow more

Introducing ml-intern, the agent that just automated the post-training... team Hugging Face It's an open-source implementation of the real research loop that our ML researchers do every day. You give it a prompt, it researches papers, goes through citations, implements ideas in GPU sandboxes, iterates and builds deeply research-backed models for any use case. All built on the Hugging Face ecosystem. It can pull off crazy things: We made it train the best model for scientific reasoning. It went through citations from the official benchmark paper. Found OpenScience and NemoTron-CrossThink, added 7 difficulty-filtered dataset variants from ARC/SciQ/MMLU, and ran 12 SFT runs on Qwen3-1.7B. This pushed the score 10% → 32% on GPQA in under 10h. Claude Code's best: 22.99%. In healthcare settings it inspected available datasets, concluded they were too low quality, and wrote a script to generate 1100 synthetic data points from scratch for emergencies, hedging, multilingual etc. Then upsampled 50x for training. Beat Codex on HealthBench by 60%. For competitive mathematics, it wrote a full GRPO script, launched training with A100 GPUs on watched rewards claim and then collapse, and ran ablations until it succeeded. All fully backed by papers, autonomously. How it works? ml-intern makes full use of the HF ecosystem: - finds papers on arxiv and reads them fully, walks citation graphs, pulls datasets referenced in methodology sections and on - browses the Hub, reads recent docs, inspects datasets and reformats them before training so it doesn't waste GPU hours on bad data - launches training jobs on HF Jobs if no local GPUs are available, monitors runs, reads its own eval outputs, diagnoses failures, retrains ml-intern deeply embodies how researchers work and think. It knows how data should look like and what good models feel like. Releasing it today as a CLI and a web app you can use from your phone/desktop. CLI: Web + mobile: And the best part? We also provisioned 1k$ GPU resources and Anthropic credits for the quickest among you to use.show more

Agent Skills are now available in Cursor. Skills let... agents discover and run specialized prompts and code.show more

🔹Comparing AI video results: • Gen-3 • Kling AI... 1.5 • Hailuo MiniMax • Luma Dream Machine I used a FLUX LoRA trained on my selfies as the input. Each model ran the same prompt 3 times, and I chose the best results: "Tracking shot of a curator standing and speaking to the camera as she welcomes museum guests. Warm, inviting atmosphere." Each model interpreted the prompt differently, and the backgrounds are interesting:show more

Qwen Image Multiple Angles LoRA is an exquisitely trained... LoRA! 📐˚₊‧꒰ა Keep character and scenes consistent, and flies the camera around! Open source got there! One of the best LoRAs I've come across lately 🙌show more

⚡️ We just added support for Nvidia G4 (RTX... Pro 6000 Blackwell Server Edition) GPUs in Colaboratory! 🔥 With a peak rating of 960 BF16 TFLOPs (~50% more than the A100-80G) and 96 GB of VRAM (20% more than the A100-80G), it’s the most efficient high-performance GPU we’ve ever released. ⏱️Time to make GPUs go brrrr 🏎️show more

🧮Turn single image into a custom LoRA model in... one click ! Should work for character and product ! This ComfyUI workflow: → Uses Gemini AI to generate 20 diverse prompts from your image → Creates 20 consistent variations with FLUX.1 Kontext → Automatically builds the dataset + trains the LoRA One image in → Trained LoRA out 🎯 #ComfyUI #LoRA #AIArt #FLUX #AutomatedAI ComfyUI Black Forest Labs 🔗 Check it out: This workflow was made for the hackathon organized by ComfyUI in SF yesterdayshow more

NVIDIA + Stanford just dropped NitroGen "plays-any-game" AI trained... on 40,000 hours of gameplay across 1,000+ games. They're releasing the dataset + model weights. 👇show more

We took a 30B model and split it in... two to write tokens in parallel instead of one at a time. Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch. We found it kept 98.7% of the original model’s quality at 2.42× faster generation.show more

✨Just released: NVIDIA Nemotron post-training multi-lingual dataset ✨ We... expanded permissive post-training dataset with addition of synthetically translated reasoning traces. ✋ Five new languages 💪 World class reasoning traces 🤗 Learn more: 📥 Download:show more

Meet Stable Audio 3.0, the open-weight model family built... for artistic experimentation. This is our open invitation to experiment with generative audio. We believe the best innovations are still waiting to be built. The 4-1-1 on 3.0: 📣 You own your outputs, and can distribute and commercialize them under the Stability AI Community License (up to $1 million in revenue). 🎵 New and improved capabilities include variable-length generation up to six minutes, and full song composition on portable devices, no GPU required. ✅ Trained on a fully licensed dataset. 🎨 You can customize the models on your own library with support for LoRa training, which we’ve documented for the first time. More on the models 👇show more

Democrats are freaking out about this video where a... TSA agent was spotted training an ICE agent on how to operate the biometrics scanner. Prior TSA female claims, "I had to take 20 exams and weeks of rigorous training". Girl, these legends learned it in less than a day. 🤣show more

Amazon Web Services CEO Matt Garman said today there... is so much demand they have NEVER retired an Nvidia A100 GPU. The A100 is nearly SIX YEARS OLD. And you're still bearish? "Because there is so much more demand than supply, there typically still is demand for the older chips .. we actually are completely sold out and have never retired an A100 server"show more

NVIDIA just launched NVIDIA App Beta. It's going to... be a replacement for NVIDIA Control Panel and GeForce Experience App. - NVIDIA app install in half the time, deliver a 50% more responsive UI, and occupy 17% less disk space than GeForce Experience.show more

Want to bring open-source vision language models to the... edge? 💻 Check out our Hugging Face article on deploying NVIDIA Cosmos Reasoning 2B across the NVIDIA Jetson family with vLLM and a Live VLM WebUI. 📖show more

This is how you unlock the next billion software... developers. The new Replit ⠕ Agent 3 (they just launched) is the most advanced vibe-coding agent in the world. 1. Smarter than any other vibe-coding model (10x more autonomous than the previous version). 2. It thinks harder and lasts longer than any other model (up to 200 minutes running fully autonomously). 3. The agent can now use an actual browser to test and fix its own code. 4. 3x faster and 10x more cost-effective than any other "Computer Use" for testing. 5. It can build other agents and automations to take care of repetitive tasks. Seeing the agent test the application autonomously is science fiction!show more

BREAKING 🚨: OpenAI launched a dedicated Vibe coding Codex... app for macOS. Codex also supports Skills and automations! Everyone is a developer now 👀show more

FLUX [DEV] LoRA Training in Google Colab ❗️❗️ Notebook... and code by the 🐐 Ostrisshow more

Vision AI agents are becoming a practical way to... automatically turn video data from the physical world into operational intelligence in factories, cities, warehouses and transportation systems. See how Roboflow, DeepHow, and Linker Vision use NVIDIA Cosmos and NVIDIA Metropolis to generate synthetic data, fine-tune vision models faster, and deploy vision agents at scale. 📰 Read the blog:show more