NVIDIA AI

Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model built for long-running agents. It delivers 5x faster inference and lowers the cost of complex agentic tasks by up to 30% versus other open frontier models.

@ollama @Nasdaq @Theoryvc @ttunguz 🙌

Ollama is back in NYC today! It's exciting to see more people and businesses realize the benefits of open models! Ownership. Affordable. Private. Thank you @Nasdaq and the @Theoryvc team.

@Mayhem4Markets 🤝

How to Run RL Autoresearch with Agent Skills | Nemotron Labs x.com/i/broadcasts/1…

Read the technical walkthrough 👇 nvda.ws/3T3cRpl

🚦 The zero-shot model misses a visible traffic signal. After post-training Cosmos 3 Nano with LoRA, it correctly identifies the intersection, and overall WTS validation accuracy increases from 54.41% to 87.14%. 📈 With NVIDIA TAO AutoML, accuracy reaches 93.35%. All in under a day.

If you want to try it out for yourself, check out the video below or follow along here: nvda.ws/4wIdgMz

We gave a coding agent a goal and a time budget: build a training environment and teach a vision model to count colored stars. Using autoresearch with NeMo RL, NeMo Gym, and reusable skills, the agent set up, trained and evaluated the model while the researcher steered the work. Qwen3-VL-2B went from 25% to 96.9% accuracy, and the agent even proposed the next experiment on its own.

@JamesMcPherson 🙌

I hear we are posting our @NVIDIAAI clusters this morning, so here we go: 4X GB10's with CRS804 RoCE for the fabric. 10GBE backhaul to the NAS in the other rack. All in a 12U 10" rack. SOTA AI, in the corner of my den. Currently serving GLM-5.2 NVFP4

Proud to support the open source community. Thanks for the Nemotron shoutout @jmorgan! 🙌

@jmorgan Thanks for the Nemotron shoutout! 🙌

Great to be on TBPN to share more on Ollama's Series B fundraise and how open models are becoming the dominant force in enterprise. Thanks for having me ☺️.

As AI models continue to grow in scale and capability, shaping a model matters just as much as its size. We're introducing a new series on AI Model Co-Design exploring the synergy between models and hardware. The first post focuses on how model dimensions influence GPU performance, and how the right design choices improve both system throughput and per-user responsiveness. You can read it here: nvda.ws/452Idiy

@Teknium 💚

Hey everyone. I haven't been very responsive on here the last week. My dog, Link, who I've raised since he was a puppy over the last 13 years, passed away yesterday after being in the vet ER's ICU since last Wednesday for heart failure. I put together some of my favorite pics of him to share so you all can see the most awesome animal friend I could ask for. I'll be a bit slow probably through this week too, hope you all can understand 🙏

And huge thank you @swyx and AIEWF for providing the platform!

@dee_hw 🙌

It's Saturday. A good day to build your own Personal AI Computer. Adding another 4x RTX PRO 6000 rig for the team.

@mervenoyann 💚

Epic panel - let's run it back soon.

@googlegemma 🦾💪

Hugging Face Gemma Challenge results are in! 📈 Over 6 days, more than 100 AI agents and humans collaborated to make Gemma 4 inference 5x faster on a single NVIDIA A10G GPU. - Fastest result: 491.8 TPS (fastest overall, but resulted in a drop in model quality in other areas) - Fastest lossless: 315 TPS A great example of what humans and agents can achieve when they work together.

@johnciannello @nvidia Let's gooo! build.nvidia.com/spark

@sgl_project 🔥

🎉 SGLang v0.5.15 is out! We spent this cycle tuning GLM-5.2 NVFP4 for production serving, now hitting 500+ tok/s/user on 8x B300 and 450 on 4x GB300 (bs=1). We will put commands to run this at the thread below, and full technical details and instructions on a blog very soon 🫡 And we have some newly supported models: Hunyuan 3 (Hy3), Hierarchical Reasoning Model (HRM-Text), NVIDIA LocateAnything-3B, Baidu Unlimited-OCR, JoyEcho, and Qwen3.6. Here are highlights for this release: - Breakable CUDA Graph is now the default capture path - Native web search built in, powered by @ExaAILabs - Decode context parallelism for MLA models, including DeepSeek V3 - FlashInfer all-to-all for routed MoE - DeepSeek-V4: FlashMLA sparse prefill now on by default (>10% throughput on long context), plus a non-paged indexer for long-context prefill (>5% e2e) We welcomed 43 new contributors, and thanks again for our amazing partners and model makers: @NVIDIAAI @AMD @intel @Zai_org @TencentHunyuan @Alibaba_Qwen @deepseek_ai @Sapient_Int Now. MAX LOAD! MAX OUTPUT! 🚀