Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation

Zhuolin Yang, Zihan Liu, Yang Chen, Wenliang Dai, Boxin Wang, Sheng-Chieh Lin, Chankyu Lee, Yangyi Chen, Dongfu Jiang, Jiafan He, Renjie Pi, Grace Lam, Nayeon Lee, Alexander Bukharin, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping

Abstract

We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance in the 2025 International Mathematical Olympiad (IMO), the International Olympiad in Informatics (IOI), and the ICPC World Finals, demonstrating remarkably high intelligence density with 20x fewer parameters. In contrast to Nemotron-Cascade 1, the key technical advancements are as follows. After SFT on a meticulously curated dataset, we substantially expand Cascade RL to cover a much broader spectrum of reasoning and agentic domains. Furthermore, we introduce multi-domain on-policy distillation from the strongest intermediate teacher models for each domain throughout the Cascade RL process, allowing us to efficiently recover benchmark regressions and sustain strong performance gains along the way. We release the collection of model checkpoint and training data.

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation

Abstract

We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance in the 2025 International Mathematical Olympiad (IMO), the International Olympiad in Informatics (IOI), and the ICPC World Finals, demonstrating remarkably high intelligence density with 20x fewer parameters. In contrast to Nemotron-Cascade 1, the key technical advancements are as follows. After SFT on a meticulously curated dataset, we substantially expand Cascade RL to cover a much broader spectrum of reasoning and agentic domains. Furthermore, we introduce multi-domain on-policy distillation from the strongest intermediate teacher models for each domain throughout the Cascade RL process, allowing us to efficiently recover benchmark regressions and sustain strong performance gains along the way. We release the collection of model checkpoint and training data.
Paper Structure (50 sections, 4 equations, 5 figures, 6 tables)

This paper contains 50 sections, 4 equations, 5 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Main Results
  3. Supervised Fine-Tuning
  4. Training Framework
  5. Overview
  6. Chat Template
  7. SFT Data Curation
  8. Math
  9. Code Reasoning
  10. Science
  11. Long Context
  12. General Chat
  13. Instruction Following
  14. Safety
  15. Conversational Agent
  16. ...and 35 more sections

Figures (5)

  • Figure 1: (Left) The chat template uses adjacent <think></think> tokens to indicate non-thinking mode, and a single <think> followed by \\ n to indicate thinking mode. (Right) For tool calling, the available tools are listed in the system prompt. The model is instructed to call tools within the <tool_call> and </tool_call> tags.
  • Figure 2: Nemotron-Cascade 2 applies Cascade RL with the sequential, domain-wise ordering after SFT, leading to substantial improvements across the corresponding domains.
  • Figure 3: Training dynamics and downstream evaluation.
  • Figure 4: IMO-ProofBench (Advanced) score graded by LLM ProofAutoGrader (DeepSeek-V3.2-Speciale).
  • Figure :