Nemotron 3 Ultra performed GPT 5.5 level 10× cheaper We gave three same prompts to build HTML5 canvas with real physics. At first scene we have water in a spinning drum. Galton board - balls through pegs into bins. And a block collision setup with extreme mass differences. Outputs: Nemotron 3 Ultra: 11.3k tokens, $0.051 GPT 5.5: 11.0k tokens, $0.57 Nemotron stays right on GPT 5.5's heels, but at 10× cheaper. The gap in quality is far smaller than the gap in price.

nemotron 3 nano omni-30B reasoning at Q8 running autonomously on my dgx spark right now. 58 tok/s. 1 million context. multimodal. hermes agent is using it to research xAI's new grok algorithm that dropped yesterday. pulling repos. scanning code. breaking it down. all while i post this. 30B model. 58 tokens per second. 1M context. reads images and video. locally. for free. nobody is talking about this model and that's insane.

NVIDIA Nemotron 3.5 Lightning is now live on OpenRouter. A 30B hybrid MoE with 3B active params, distilled from Nemotron 3 Ultra. Built for high-volume, specialized AI agent workloads, delivering up to 4× higher throughput and up to 30% faster task completion compared to similar models in its class.

NVIDIA Nemotron 3 Ultra now available in Command Code! Strongest US open model yet! 🍀 • 1M context • 5x faster inference • 550B MoE frontier-intelligence open model DEAL 2.3x usage 🎟️ $1 Go plan gets you ~$23 usage on Nemotron Woah, it's fast x taste compliance is great!

I tested @NVIDIAAI's Nemotron Puzzle 75B locally on my GX10 with 3 HTML game challenges: - Tower Defense 🗼(~ 1m30 at 60 tok/s) - Wrecking ball 🏐(~ 40 s at 60 tok/s) - 2D Roguelite 👾(~ 1m at 60 tok/s) The results are satisfying considering that there was no harness involved (ignoring the shaggy physics of the wrecking ball one)

Benchmarking @NVIDIAAI's Nemotron Puzzle 75B locally on the GX10. NVFP4 via vLLM's OpenAI API, MTP speculative decoding, forced 1,500-token generations. 🏃‍♀️22.75 tok/s in a single session into 88.85 cumulative at 7 sessions. 🧍Baseline without MTP: ~16.3. Scripts + setup:

本周两篇： Nvidia: Nemotron-Labs-TwoTower: Diffusion Language Modeling with Pretrained Autoregressive Context ByteDance Seed: EdgeBench: Unveiling Scaling Laws of Learning from Real-World Environments