Topic
Nemotron 3 Ultra
2 episodes
-
Tuning the harness, not the model: a Nemotron 3 Ultra playbook
A LangChain/NVIDIA case study claims harness tuning alone can push Nemotron 3 Ultra to 0.86 on Deep Agents at ~$4.48/run vs $43.48 for Opus 4.8, with parity latency. The hosts parse the real mechanism (matched scaffolding vs post-training), test limits (where harness hits a ceiling), and weigh who actually benefits. They surface concrete repos (langchain-ai/deepagents, langchain-ai/deepagentsjs) and a vendor profile workflow, then poke at the article’s reliance on Deep Agents and the cost math. Final take: a plausible win for teams already deep into harness work, not a universal unlock, with the hosts pushing back on ‘ten-x cheaper’ framing and under-specified benchmark footnotes.
-
Nemotron 3 Ultra: Open, Efficient Mixture of Experts Hybrid Mamba Transformer Model for Agentic Reasoning
Nemotron 3 Ultra is NVIDIA's 550B-parameter Mixture-of-Experts hybrid Mamba-Attention model with 55B active parameters per token, pre-trained on 20 trillion tokens and extended to 1M context. It achieves 6× higher inference throughput than comparable open models while maintaining on-par accuracy, using LatentMoE, Multi-Token Prediction, NVFP4 low-precision training, and multi-teacher on-policy distillation. The entire model, training recipes, and datasets are open-sourced on HuggingFace.