NVIDIA AI's banner
NVIDIA AI's profile picture

NVIDIA AI

@NVIDIAAI • 351,243 subscribers

Teaching your AI new tricks.

Shorts

A speech model can support Arabic and still struggle with local dialects. Fine-tuning NVIDIA Nemotron 3.5 ASR cut word error rate on Najdi and Hijazi Saudi Arabic from 55% to 30%. Check out our new tutorial and learn how to adapt it for other dialects and languages:

A speech model can support Arabic and still struggle with local dialects. Fine-tuning NVIDIA Nemotron 3.5 ASR cut word error rate on Najdi and Hijazi Saudi Arabic from 55% to 30%. Check out our new tutorial and learn how to adapt it for other dialects and languages:

678,753 views

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.

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.

763,029 views

Barcelona showed up ready to build 🦞. The road to #NVIDIAGTC Berlin is officially underway.

Barcelona showed up ready to build 🦞. The road to #NVIDIAGTC Berlin is officially underway.

20,945 views

Build-A-Claw London ✅️ A visit from Jensen Huang ✅️ A very good day to have a raffle ticket ✅️ Thank you to everyone who stopped by! See you in the next city.

Build-A-Claw London ✅️ A visit from Jensen Huang ✅️ A very good day to have a raffle ticket ✅️ Thank you to everyone who stopped by! See you in the next city.

14,700 views

This #CVPR2026 paper from our research team is trending #1 on Hugging Face 🤗 Meet LocateAnything: a vision-language detection model that rethinks bounding box prediction. For AI agents and robots, “seeing” is only useful if a model can pinpoint where something is fast enough to act. Trained on 138M high-quality samples, LocateAnything decodes bounding boxes in parallel instead of one coordinate at a time, improving localization accuracy while dramatically increasing throughput for visual grounding and detection. Project page:

This #CVPR2026 paper from our research team is trending #1 on Hugging Face 🤗 Meet LocateAnything: a vision-language detection model that rethinks bounding box prediction. For AI agents and robots, “seeing” is only useful if a model can pinpoint where something is fast enough to act. Trained on 138M high-quality samples, LocateAnything decodes bounding boxes in parallel instead of one coordinate at a time, improving localization accuracy while dramatically increasing throughput for visual grounding and detection. Project page:

341,842 views

Happy Friday! We just put DeepSeek-V4-Pro up on It’s the world’s largest open source model at 1.6T parameters, and you can run it for free running on NVIDIA Blackwell GPUs. Try the NVIDIA NIM API →

Happy Friday! We just put DeepSeek-V4-Pro up on It’s the world’s largest open source model at 1.6T parameters, and you can run it for free running on NVIDIA Blackwell GPUs. Try the NVIDIA NIM API →

207,975 views

What if your model handled its own tuning? ⚙️ With NVIDIA TAO 7, just prompt your coding agent what you want in plain language and let it do the rest. • Agent skills plug into your coding agent and improve accuracy. • AutoML removes the guesswork from hyperparameters. LLM-guided tuning finds strong configs up to 2x faster. • Fine-tune any Hugging Face CV or VLM model on local NVIDIA GPUs. • Data enhanced fine-tuning helps your agent identify why the model fails, then fix it. Try TAO skills ➡️

What if your model handled its own tuning? ⚙️ With NVIDIA TAO 7, just prompt your coding agent what you want in plain language and let it do the rest. • Agent skills plug into your coding agent and improve accuracy. • AutoML removes the guesswork from hyperparameters. LLM-guided tuning finds strong configs up to 2x faster. • Fine-tune any Hugging Face CV or VLM model on local NVIDIA GPUs. • Data enhanced fine-tuning helps your agent identify why the model fails, then fix it. Try TAO skills ➡️

42,294 views

Artificial Analysis Image-to-video generation is just as impressive. Input image: "Generate a 16:9 image from a dashcam view of a formula 1 racing event" Video prompt: "A high-speed racing event where a car navigates multiple winding turns" 🔊 Sound on - generated by Cosmos 3.

Artificial Analysis Image-to-video generation is just as impressive. Input image: "Generate a 16:9 image from a dashcam view of a formula 1 racing event" Video prompt: "A high-speed racing event where a car navigates multiple winding turns" 🔊 Sound on - generated by Cosmos 3.

49,260 views

“Every enterprise needs a claw strategy.” How did LangChain go from a weekend project to 1B+ downloads in 3 years? We sat down with CEO and co-founder Harrison Chase (Harrison Chase) to talk deep agents, evolving agent architectures, and what’s coming next. 🎧 Full episode:

“Every enterprise needs a claw strategy.” How did LangChain go from a weekend project to 1B+ downloads in 3 years? We sat down with CEO and co-founder Harrison Chase (Harrison Chase) to talk deep agents, evolving agent architectures, and what’s coming next. 🎧 Full episode:

47,817 views

Selected as a best paper finalist at #CVPR2026: PixelDiT from NVIDIA Research In most image generation models, a pretrained autoencoder compresses the image before any diffusion happens, causing quality loss that accumulates across the entire pipeline. PixelDiT, or Pixel Diffusion Transformers, removes this step entirely. It's a single-stage model that learns the diffusion process directly in pixel space, end-to-end.

Selected as a best paper finalist at #CVPR2026: PixelDiT from NVIDIA Research In most image generation models, a pretrained autoencoder compresses the image before any diffusion happens, causing quality loss that accumulates across the entire pipeline. PixelDiT, or Pixel Diffusion Transformers, removes this step entirely. It's a single-stage model that learns the diffusion process directly in pixel space, end-to-end.

28,504 views

Factories are getting a new AI brain 🧠 Introducing NVIDIA Factory Operations Blueprint (FOX), a reference design for building factory manager agents that monitor operations, reason across real-time data, and coordinate specialized AI agents to help resolve issues at scale. Early adopters including Hon Hai Technology Group (Foxconn), Pegatron, Advantech USA, and Wistron AiEdge Corporation are already seeing gains in productivity, quality, and efficiency. Read more: #NVIDIAGTC

Factories are getting a new AI brain 🧠 Introducing NVIDIA Factory Operations Blueprint (FOX), a reference design for building factory manager agents that monitor operations, reason across real-time data, and coordinate specialized AI agents to help resolve issues at scale. Early adopters including Hon Hai Technology Group (Foxconn), Pegatron, Advantech USA, and Wistron AiEdge Corporation are already seeing gains in productivity, quality, and efficiency. Read more: #NVIDIAGTC

25,781 views

With 3M+ downloads and counting, NVIDIA Cosmos is redefining physical AI. Announced at #CORL25, new Cosmos updates are allowing developers to generate diverse data for accelerating training robot models at scale. 👏 Cosmos Predict 2.5 will combine three models into one powerful model—reducing complexity, powering up to 30s video generation, and enabling multi-view simulations. 👏 Cosmos Transfer 2.5 will be 3.5x smaller yet faster and sharper—generating photorealistic synthetic data from 3D scenes or spatial inputs. 🔗

With 3M+ downloads and counting, NVIDIA Cosmos is redefining physical AI. Announced at #CORL25, new Cosmos updates are allowing developers to generate diverse data for accelerating training robot models at scale. 👏 Cosmos Predict 2.5 will combine three models into one powerful model—reducing complexity, powering up to 30s video generation, and enabling multi-view simulations. 👏 Cosmos Transfer 2.5 will be 3.5x smaller yet faster and sharper—generating photorealistic synthetic data from 3D scenes or spatial inputs. 🔗

40,392 views

The next wave of AI agents won’t just read text—they’ll reason over video and extract actionable insights.💡 In this #NVIDIAGTC session, be the first to hear about the new features powering visual AI agents with NVIDIA Blueprint for Video Search and Summarization (VSS), open VLMs including the latest Cosmos Reason, for development across edge, on‑prem, and cloud workloads. Add to schedule 👉 📅 Tues, March 17 | 10:00 a.m PT Speakers🎤: Roopa Prabhu and Adam Ryason

The next wave of AI agents won’t just read text—they’ll reason over video and extract actionable insights.💡 In this #NVIDIAGTC session, be the first to hear about the new features powering visual AI agents with NVIDIA Blueprint for Video Search and Summarization (VSS), open VLMs including the latest Cosmos Reason, for development across edge, on‑prem, and cloud workloads. Add to schedule 👉 📅 Tues, March 17 | 10:00 a.m PT Speakers🎤: Roopa Prabhu and Adam Ryason

12,850 views

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

Meet NVIDIA Nemotron 3.5 Super VL Our new open model combines reasoning and tool use with text, image, document, and video understanding for long-running AI agents. 120B total parameters with 12B active 🧵
1:20
NVIDIAAI's profile picture

Meet NVIDIA Nemotron 3.5 Super VL Our new open model combines reasoning and tool use with text, image, document, and video understanding for long-running AI agents. 120B total parameters with 12B active 🧵

NVIDIA AI

548,648 views • 2 days ago

Big shoutout to all of our partners for making Nemotron 3.5 Super VL their own through post-training 🙌 Here’s what they’ve built 🧵
0:20
NVIDIAAI's profile picture

Big shoutout to all of our partners for making Nemotron 3.5 Super VL their own through post-training 🙌 Here’s what they’ve built 🧵

NVIDIA AI

36,071 views • 1 day ago

When several people talk at once, a transcript can get messy fast. Our new Nemotron 3 Diarization model tracks who spoke when, even when voices overlap. It handles up to eight speakers, has 100M parameters, and is now available on Hugging Face 🤗
0:35
NVIDIAAI's profile picture

When several people talk at once, a transcript can get messy fast. Our new Nemotron 3 Diarization model tracks who spoke when, even when voices overlap. It handles up to eight speakers, has 100M parameters, and is now available on Hugging Face 🤗

NVIDIA AI

1,065,688 views • 24 days ago

Agents may need to read documents, interpret charts, and search the web to complete a task. Nemotron 3.5 Super VL brings text and visual understanding into one model, reducing the need to pass context between separate text and vision models.
3:01
NVIDIAAI's profile picture

Agents may need to read documents, interpret charts, and search the web to complete a task. Nemotron 3.5 Super VL brings text and visual understanding into one model, reducing the need to pass context between separate text and vision models.

NVIDIA AI

26,250 views • 2 days ago

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.
2:59
NVIDIAAI's profile picture

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.

NVIDIA AI

1,330,976 views • 4 months ago

Your agent got the right answer. How much work did it take to get there? We worked with Nous Research/Teknium 🪽 on a hands-on walkthrough of NVIDIA NeMo Relay to collect traces for Hermes Agent. Run two example scenarios, follow the agent’s calls and retries, and see the full trace in Arize Phoenix. The blog also looks at how Nous used traces and task results to evaluate fixes across repeated runs. TechBlog: 📽️ by our own pico:
3:50
NVIDIAAI's profile picture

Your agent got the right answer. How much work did it take to get there? We worked with Nous Research/Teknium 🪽 on a hands-on walkthrough of NVIDIA NeMo Relay to collect traces for Hermes Agent. Run two example scenarios, follow the agent’s calls and retries, and see the full trace in Arize Phoenix. The blog also looks at how Nous used traces and task results to evaluate fixes across repeated runs. TechBlog: 📽️ by our own pico:

NVIDIA AI

86,376 views • 17 days ago

An AI agent makes a mistake early in a task, then keeps going in the wrong direction. Our researchers built PivotOPD to teach agents how to avoid those mistakes and recover when they happen. During training, a teacher model shows the agent a better action and how to get back on track over the next few steps. Read the paper and watch how it works:
1:51
NVIDIAAI's profile picture

An AI agent makes a mistake early in a task, then keeps going in the wrong direction. Our researchers built PivotOPD to teach agents how to avoid those mistakes and recover when they happen. During training, a teacher model shows the agent a better action and how to get back on track over the next few steps. Read the paper and watch how it works:

NVIDIA AI

38,156 views • 10 days ago

Introducing Cosmos 3: Our latest frontier model for Physical AI Cosmos 3 is the world’s first fully open omnimodel with native vision reasoning, world and action generation. Today we’re releasing Super (32B) and Nano (8B) variants.
3:13
NVIDIAAI's profile picture

Introducing Cosmos 3: Our latest frontier model for Physical AI Cosmos 3 is the world’s first fully open omnimodel with native vision reasoning, world and action generation. Today we’re releasing Super (32B) and Nano (8B) variants.

NVIDIA AI

429,693 views • 4 months ago

Nemotron 3 Ultra is coming this week. ⌛️
2:11
NVIDIAAI's profile picture

Nemotron 3 Ultra is coming this week. ⌛️

NVIDIA AI

396,179 views • 4 months ago

Meet Nemotron 3 Nano Omni 👋 Our latest addition to the Nemotron family is the highest efficiency, open multimodal model with leading accuracy. 30B parameters. 256K context length. 🧵👇
4:24
NVIDIAAI's profile picture

Meet Nemotron 3 Nano Omni 👋 Our latest addition to the Nemotron family is the highest efficiency, open multimodal model with leading accuracy. 30B parameters. 256K context length. 🧵👇

NVIDIA AI

461,004 views • 5 months ago

Introducing Cosmos 3 Edge: our open frontier world model built to run on-device. Cosmos 3 Edge helps robots learn and act, autonomous vehicles understand road scenes and predict intent, and vision AI agents reason across live video for smart infrastructure. With 4B parameters and a 2B Nemotron-based reasoner, you can run it on DGX Spark, NVIDIA Jetson, and more.
0:22
NVIDIAAI's profile picture

Introducing Cosmos 3 Edge: our open frontier world model built to run on-device. Cosmos 3 Edge helps robots learn and act, autonomous vehicles understand road scenes and predict intent, and vision AI agents reason across live video for smart infrastructure. With 4B parameters and a 2B Nemotron-based reasoner, you can run it on DGX Spark, NVIDIA Jetson, and more.

NVIDIA AI

212,263 views • 2 months ago

Nemotron 3 Diarization ranked #1 of 12 systems in Voice Arena's initial Diarization-Bench results. Its 14.72% error rate was ~24% lower than the runner-up. Here’s a four-speaker comparison from the benchmark. Full results:
0:31
NVIDIAAI's profile picture

Nemotron 3 Diarization ranked #1 of 12 systems in Voice Arena's initial Diarization-Bench results. Its 14.72% error rate was ~24% lower than the runner-up. Here’s a four-speaker comparison from the benchmark. Full results:

NVIDIA AI

54,556 views • 24 days ago

Most motion papers tailor one controller to one specific task. This year at SIGGRAPH, our research team asks: can motor control itself be pretrained and reused? Generative Pretrained Controllers, or GPC, turn motor skills into a vocabulary of discrete tokens and train a transformer-based generative controller through next-token prediction. Just like GPT, the same pretrained controller can then be fine-tuned to solve new tasks. Trained on 600+ hours of motion, GPC runs in real-time inside a physics simulation, producing natural and physically grounded behaviors for interactive control.
1:45
NVIDIAAI's profile picture

Most motion papers tailor one controller to one specific task. This year at SIGGRAPH, our research team asks: can motor control itself be pretrained and reused? Generative Pretrained Controllers, or GPC, turn motor skills into a vocabulary of discrete tokens and train a transformer-based generative controller through next-token prediction. Just like GPT, the same pretrained controller can then be fine-tuned to solve new tasks. Trained on 600+ hours of motion, GPC runs in real-time inside a physics simulation, producing natural and physically grounded behaviors for interactive control.

NVIDIA AI

176,369 views • 3 months ago

A video world model can make a convincing clip and still get the physics wrong. Our researchers just released Physis-Lang, an open self-evolving framework that adds physics reasoning to video captions. The captions explain why and how a scene unfolds. We use them to fine-tune world models and add that reasoning to prompts when generating video. Adding physics reasoning to the prompt alone improved NVIDIA Cosmos 3’s PhyGenBench score by 5.62 points without retraining. Paper and project:
2:01
NVIDIAAI's profile picture

A video world model can make a convincing clip and still get the physics wrong. Our researchers just released Physis-Lang, an open self-evolving framework that adds physics reasoning to video captions. The captions explain why and how a scene unfolds. We use them to fine-tune world models and add that reasoning to prompts when generating video. Adding physics reasoning to the prompt alone improved NVIDIA Cosmos 3’s PhyGenBench score by 5.62 points without retraining. Paper and project:

NVIDIA AI

31,266 views • 18 days ago

When it comes to Chess, Vidit Gujrathi is an absolute legend. Recently, he entered the Hermes Agent Accelerated Business Hackathon to see who could build the best agent using Nemotron. Here's what he submitted. Special thanks to Stripe & Nous Research 🤝
1:28
NVIDIAAI's profile picture

When it comes to Chess, Vidit Gujrathi is an absolute legend. Recently, he entered the Hermes Agent Accelerated Business Hackathon to see who could build the best agent using Nemotron. Here's what he submitted. Special thanks to Stripe & Nous Research 🤝

NVIDIA AI

156,029 views • 3 months ago

One prompt built and deployed a visual AI agent for a manufacturing line in under 30 minutes, with alerts, video search, and incident reports. Check out our new tutorial that shows you how to build one with the new Build Vision AI skill in the NVIDIA VSS Blueprint 3.3. Tutorial:
1:00
NVIDIAAI's profile picture

One prompt built and deployed a visual AI agent for a manufacturing line in under 30 minutes, with alerts, video search, and incident reports. Check out our new tutorial that shows you how to build one with the new Build Vision AI skill in the NVIDIA VSS Blueprint 3.3. Tutorial:

NVIDIA AI

27,790 views • 17 days ago

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:
0:29
NVIDIAAI's profile picture

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:

NVIDIA AI

155,804 views • 3 months ago

Huge thank you to everyone who downloaded Nemotron 3 Diarization and helped it trend on Hugging Face! And we appreciate all the comments. sophie a answered a few of your questions:
0:57
NVIDIAAI's profile picture

Huge thank you to everyone who downloaded Nemotron 3 Diarization and helped it trend on Hugging Face! And we appreciate all the comments. sophie a answered a few of your questions:

NVIDIA AI

21,868 views • 14 days ago

NVIDIA Metropolis Blueprint for video search and summarization (VSS) 3 is here. Now your coding agent can analyze massive live streams and libraries of videos with a simple natural language prompt. Here's what's new: - 16 new agent skills: Search, summarize, alert, report, review clips. All from natural language prompts. - One unified open source repo: Source code, Docker and Helm deployment profiles for fast, easy deployment. - Multi-video reports and Nemotron 3 Nano Omni: Insights across video and audio at scale. - 3D multi-camera tracking: Production ready + #1 SOTA for smarter scene understanding. Try VSS skills 👉
1:23
NVIDIAAI's profile picture

NVIDIA Metropolis Blueprint for video search and summarization (VSS) 3 is here. Now your coding agent can analyze massive live streams and libraries of videos with a simple natural language prompt. Here's what's new: - 16 new agent skills: Search, summarize, alert, report, review clips. All from natural language prompts. - One unified open source repo: Source code, Docker and Helm deployment profiles for fast, easy deployment. - Multi-video reports and Nemotron 3 Nano Omni: Insights across video and audio at scale. - 3D multi-camera tracking: Production ready + #1 SOTA for smarter scene understanding. Try VSS skills 👉

NVIDIA AI

126,067 views • 3 months ago

Images rarely show an object’s full 3D geometry. At #ECCV2026, our research team introduced Axolotl3D, a multimodal and occlusion-aware 3D generation model. It combines images, camera data and partial geometry to reconstruct missing regions while preserving observed ones, achieving state-of-the-art results across single- and multi-view settings. Project page:
3:17
NVIDIAAI's profile picture

Images rarely show an object’s full 3D geometry. At #ECCV2026, our research team introduced Axolotl3D, a multimodal and occlusion-aware 3D generation model. It combines images, camera data and partial geometry to reconstruct missing regions while preserving observed ones, achieving state-of-the-art results across single- and multi-view settings. Project page:

NVIDIA AI

30,593 views • 1 month ago