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.

@clipitdev super helpful agent

Hackathon entry: Clipit, an agent-first, video-forward social media tool. Clipit helps agents turn long-form content into short form clips and distribute them across social platforms. Clipit plugs into the agent environments people already use. The workflow is simple: Give the agent a video, stream, podcast, or content library Clipit digests the source and finds usable moments The agent generates clips, captions, thumbnails, reframes, and post copy Clipit queues or publishes the final assets The human stays in the loop for taste, approval, and brand judgment The bigger idea: agents need tools that let them operate, not just answer. Clipit is the creation, editing, and posting layer for short-form video. We are building the searchable memory layer behind it, helping agents find old material, resurface context, and keep a brand’s video history useful. Together, they give agents a real media workflow: find content, generate clips, edit them, and distribute them. Clipit is the video production layer for AI agents. @NousResearch @NVIDIAAI @stripe

@jetbrains @huggingface 🥳

Join us at #SIGGRAPH2026 for more papers, breakthroughs, and sessions. nvda.ws/4vtIiHQ

3D scene reconstruction works great until the camera never sees part of the scene. ArtiFixer from NVIDIA Research is an open autoregressive model that fills in the missing geometry that other methods leave blank. #SIGGRAPH2026 paper, code + demo: nvda.ws/4oILqNd

@peholderrieth @nvidia Welcome!

Hi everyone! I’ve moved to the Bay Area for a summer research internship at @nvidia. Beyond exciting work, I'd love to meet new people doing exciting stuff (incl. stuff I don't work on myself rn!). If you’re around, I’d love to connect! Even if just for a jam session!

@calebfoundry 🔥

my first ever DGX Spark.. i will do my best to use it to share deeper knowledge around local models, inference, agents, coding, etc. thank you @NVIDIAAI

@physical_int @nvidia 🦾

New work with @nvidia: evaluating robot policies entirely inside a world model. The policy acts, the model imagines the consequences, and the imagined evals predict real-world results. 🧵 real vs world-model rollout side by side📷

@DevRico003 Agree

@MichaelGannotti @nvidia 🦾

And…. IT HAS ARRIVED!!!!! My first @NVIDIAAI DGX Spark with 128 gb Unified Memory and a 4 tb ssd. To say I’m excited would be … and understatement! Can’t wait to load up some @nvidia LLMs and start rocking!

16x parallel Gemma-4-26B-A4B-NVFP4 runs 🤯🤯🤯 18 output tokens/s, aggregate 300 tok/s 🫪 1 DGX Spark with 128 GB unified memory Concurrency so high I had to demo it programmatically It can go up to 32 even! 🤯 But then my screen would not have been readable for you And this is not even using flashinfer yet! Please reply if you know whether support is on the way Note that this is not dumb e4b or e2b that you can run on the average laptop. This is the big Gemma MoE Model link: huggingface.co/nvidia/Gemma-4…

@leyten @Zai_org 🙌

Wow, it has happened! 30.55 tok/s on GLM-5.2 4-bit (from @Zai_org) ran by six RTX Pro 6000's across the USA scattered over WAN! I can't believe this. It was an insane build, you can read more about it on github.com/leyten/shard

@Tech2Wild @nvidia 💚

@MatsuoInstitute 🦾

松尾研究所でDGX Sparkを2台導入しました。 こちらが、DGX SparkとLLM社員の2ショット写真です

@jichiep 🙌

parakeet.cpp transcription running on a Raspberry Pi 5 CPU at 7x real-time on a 60s clip under thermal throttling. Of course speed and power consumption could be better on a GPU, but still, nice of @NVIDIAAI to release this model.

@svgdevop 💚

@poolsideai Congrats @poolsideai team! 🙌

Today we’re releasing the weights for Laguna M.1, our most capable model to date, with a 256K context length. Both base and post-trained checkpoints are now available on Hugging Face under Apache 2.0.

@RaidOwlTweets 🦾

The DGX Spark cluster has a new home…lots of room to grow 🤔

@itsLuisV8 @SpaceTimeViking 🔥

@SpaceTimeViking @NVIDIAAI LFG!!! so awesome. thanks for making the @NVIDIAAI spark even more awesome with your developments!

Qwen3.6 27B getting some love on the new AEON ULTIMATE VLLM image @NVIDIAAI DGX SPARK OPTIMIZED! github.com/AEON-7/Qwen3.6…

@joshuamschultz 🙌

The mini, mobile AI lab

@mervenoyann Well ... you of course need to stop by our HQ

I'm in SF starting 27th, who should I meet and which event do I join 👀