Wei Ping

🚀 Introducing Nemotron-Cascade 2 🚀 Just 3 months after Nemotron-Cascade 1, we’re releasing Nemotron-Cascade 2: an open 30B MoE with 3B active parameters, delivering best-in-class reasoning and strong agentic capabilities. 🥇 Gold Medal-level performance on IMO 2025, IOI 2025, and ICPC World Finals 2025: • Capabilities once thought achievable only by frontier proprietary models (e.g. Gemini Deep Think) or frontier-scale open models (i.e. DeepSeek-V3.2-Speciale-671B-A37B). • Remarkably high intelligence density with 20× fewer parameters. 🏆 Best-in-class across math, code reasoning, alignment, and instruction following: • Outperforms the latest Qwen3.5-35B-A3B (2026-02-24) and even larger Qwen3.5-122B-A10B (2026-03-11). 🧠 Powered by Cascade RL + multi-domain on-policy distillation: • Significantly expand Cascade RL across a much broader range of reasoning and agentic domains than Nemotron-Cascade 1, while distilling from the strongest intermediate teacher models throughout training to recover regressions and sustain gains. 🤗 Model + SFT + RL data: 👉 huggingface.co/collections/nv… 📄 Technical report: 👉 research.nvidia.com/labs/nemotron/…

I like this Nvidia RL paper for its complete reproducibility, so much so that, in Feynman's language, you can "invent" the whole RL training pipeline yourself. A ton of persuasive ablations you would find missing even in those frontier model reports. Some takeaways: - RLHF is important as a warmup stage for even math and code RL. In a lot of tech reports we see that reasoning RL is often the first stage of training. The authors compare training math and coding directly after SFT vs after RLHF. They find lifted math/coding performances after RLHF on all benchmarks and even just training RLHF shows lifts. - It is beneficial to train math RL on different context lengths progressively. The authors train it in three stages: 24K, 32K, 40K. Each stage has its own use. The first two stages are motivated by the fact that models have a high probability of going over the max context length, so training them at 24/32K stabilizes the reasoning and makes it more effective, and is judged by the decrease in incomplete ratio. The final stage is motivated by the opposite problem: when extending to longer context (eg. 64K), the model cannot use all of the context effectively to solve hard AIME problems. So the authors introduce a third stage to let models extend their comfortable context length to 40K and see lift in performance. An interesting property is that when the context size is small (<=24K), ablations show that throwing away responses going over the context size is more beneficial than assigning them 0 reward. But at a longer context size (>=32K), assigning 0 reward is more beneficial. (p2) Another interesting property is that they empirically find that training with a temperature of 1 (as opposed to 0.6/0.8) leads to better performances of math and coding, and needs to be carefully maintained to not have entropy explode (p3). - Probably the most interesting part: you can improve SWE performances with RL that does not execute code in an environment at all. Specifically, to solve the code repair problem of a SWE task, the training setup is as simple as giving the model the problem files (with some noise) and asking it to come up with a patch. Since there is no code execution to give an outcome reward, the reward, which is very novel, is to let an LLM compare the predicted patch with the true patch, and get the probability of the model outputting the token "yes". As a probability, this reward naturally falls into [0, 1]. By training on this signal alone, the authors are able to scale up the number of tasks. And they find lift in SWE-Bench verified. - Plain old policy gradient work. The authors choose to keep the training fully on-policy. As a result, each rollout is trained for one gradient update, and therefore the importance sampling ratio is always 1. The authors state that this is for training stability and avoids entropy collapse. Overall, the pipeline is RLHF -> instruction following -> math -> coding -> SWE, and the authors keep track of benchmark performances after each stage to see the dynamics. They also go into details of data preparation, reward function, dynamic filtering for each stage. A great resource for the open source community.

Announcing NVIDIA Nemotron 3 Super! 💚120B-12A Hybrid SSM Latent MoE, designed for Blackwell 💚36 on AAIndex v4 💚up to 2.2X faster than GPT-OSS-120B in FP4 💚Open data, open recipe, open weights Models, Tech report, etc. here: research.nvidia.com/labs/nemotron/… And yes, Ultra is coming!

Nemotron 3 Super is here — 120B total / 12B active, Hybrid SSM Latent MoE, designed for Blackwell. Truly open: permissive license, open data, open training infra. See analysis on @ArtificialAnlys Details in thread 🧵below:

Introducing Nemotron-Terminal: a systematic data engineering pipeline for scaling LLM Terminal Agents. We bridge the gap between open models and proprietary models with a fully open synthetic-to-real trajectory pipeline. 🤯The payoff: SFT on our Nemotron-Terminal-Corpus boosts Qwen3-32B from 3.4% → 27.4% on Terminal-Bench 2.0 (+24.0), rivaling models multiple its size. What makes it work? 🌟Terminal-Task-Gen: A lightweight data curation pipeline that seamlessly combines the adaptation of existing datasets with robust synthetic task construction. 🌟Nemotron-Terminal-Corpus: A massive, open-source dataset covering diverse terminal interactions, which contains explicit planning and execution traces for complex long-horizon tasks. And we’re releasing everything: 📦 Nemotron-Terminal-Corpus (Large-scale dataset) 🤖 Nemotron-Terminal models (8B, 14B, 32B) Paper: arxiv.org/abs/2602.21193 HF Daily: huggingface.co/papers/2602.21… Models & Data: huggingface.co/collections/nv… Our tech report just hit the #1 spot on Hugging Face Daily Papers! We're also incredibly excited to see the open-source community putting our work to the test, with the Nemotron-Terminal-Corpus dataset currently trending at over 1,800 downloads and counting. We can't wait to see what the community build with it!

🚀 Announcing Nanbeige4.1-3B – our latest open-source 3B model mastering reasoning, preference alignment, & agent capabilities! Try it now: huggingface.co/Nanbeige/Nanbe… reddit.com/r/LocalLLaMA/c…

Cool demo samples here — check them out! 🔥 research.nvidia.com/labs/adlr/UALM/

@rogerliuty Congrats!

Feels great to see our efforts paying off!😁

@andrew_n_carr @Zhen4good Yes

@Zhen4good @_weiping Text+vision is likely lower quality than text only - I believe that's the point being made

Very enlightening results from removing vision–text SFT 👀 • Strong vision–text pretraining + text-only SFT (zero vision) already boosts visual reasoning & tool use • Then, add multimodal RL → SOTA on both vision and text • Vision–text SFT hurts generalization ; IMO, likely due to lower-diversity, lower-quality trajectories compared to text-only SFT data

Quantization-aware distillation (QAD) delivers near-BF16 accuracy with NVFP4! Amazing work led by @huizi_mao

Very enlightening results from removing vision–text SFT!! • Strong vision–text pretraining + text-only SFT (zero vision) already boosts visual reasoning & tool use • Then, add multimodal RL → SOTA on both vision and text • Vision–text SFT hurts generalization ; IMO, likely due to lower-diversity, lower-quality trajectories compared to text-only SFT data

Kimi K2.5 tech report just dropped! Quick hits: - Joint text–vision training: pretrained with 15T vision-text tokens, zero-vision SFT (text-only) to activate visual reasoning - Agent Swarm + PARL: dynamically orchestrated parallel sub-agents, up to 4.5× lower latency, 78.4% on BrowseComp - MoonViT-3D: a unified image–video encoder with 4× temporal compression, enabling 4× longer videos in the same context - Toggle: token-efficient RL, 25–30% fewer tokens with no accuracy drop Here's our work toward scalable, real-world agentic intelligence. More details in the report 👉github.com/MoonshotAI/Kim…

@Kimi_Moonshot so cool!

@jasondeanlee don’t even need to have a frontier model to be a frontier ai lab now?

What's the secret? Long vesting I bet

@wzihanw code is cheap. non-happy-path tests are still costly.

I'm seeing a ubiquitous trend where: Code is cheap; words are costly. Answers are cheap; questions are costly.

RLVR is powerful — but how do you train with multiple rewards effectively? 🤔 🎯GDPO (not GRPO) is coming. We introduce Group reward-Decoupled Normalization Policy Optimization (GDPO), a new multi-reward RL algorithm that consistently improves per-reward convergence over GRPO across a wide range of tasks. (1/n)

feels like me and my 7-yo son, with the engineer desperately holding the cute humanoid back from going feral😂

@Guangxuan_Xiao @MITEECS @thinkymachines Congrats, Dr. @Guangxuan_Xiao!

Life update: Wrapped up my PhD at @MITEECS 🎓 Super excited to start working on pre-training at @thinkymachines.

Thanks for sharing our work, @omarsar0 — excited to see the discussion it sparks!

Since the release, many have asked why cascaded, domain-wise RL is so resistant to catastrophic forgetting in Nemotron-Cascade. It really comes down to the nature of RL, the structure of the problem, and strong execution. We break it down in our report 👇