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

MOSS-TTS Technical Report

StreakMind: AI detection and analysis of satellite streaks in astronomical images with automated database integration

VibeServe: Can AI Agents Build Bespoke LLM Serving Systems?

delta-mem: Efficient Online Memory for Large Language Models

MCP-Cosmos: World Model-Augmented Agents for Complex Task Execution in MCP Environments

Beyond Reasoning: Reinforcement Learning Unlocks Parametric Knowledge in LLMs

Debiased Model-based Representations for Sample-efficient Continuous Control

Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputs

Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States

Relit-LiVE: Relight Video by Jointly Learning Environment Video

Positive Alignment: Artificial Intelligence for Human Flourishing

LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?

Unmasking On-Policy Distillation: Where It Helps, Where It Hurts, and Why

A Single Neuron Is Sufficient to Bypass Safety Alignment in Large Language Models

SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training

ELF: Embedded Language Flows

PaperFit: Vision-in-the-Loop Typesetting Optimization for Scientific Documents

Rubric-based On-policy Distillation

CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Models

TMAS: Scaling Test-Time Compute via Multi-Agent Synergy

Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs

Qwen-Image-2.0 Technical Report

MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

Learning while Deploying: Fleet-Scale Reinforcement Learning for Generalist Robot Policies

Fast Byte Latent Transformer

AI Co-Mathematician: Accelerating Mathematicians with Agentic AI

HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents

Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers

LLMs Improving LLMs: Agentic Discovery for Test-Time Scaling

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

MOSS-TTS Technical Report

StreakMind: AI detection and analysis of satellite streaks in astronomical images with automated database integration

VibeServe: Can AI Agents Build Bespoke LLM Serving Systems?

delta-mem: Efficient Online Memory for Large Language Models

MCP-Cosmos: World Model-Augmented Agents for Complex Task Execution in MCP Environments

Beyond Reasoning: Reinforcement Learning Unlocks Parametric Knowledge in LLMs

Debiased Model-based Representations for Sample-efficient Continuous Control

Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputs

Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States

Relit-LiVE: Relight Video by Jointly Learning Environment Video

Positive Alignment: Artificial Intelligence for Human Flourishing

LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?

Unmasking On-Policy Distillation: Where It Helps, Where It Hurts, and Why

A Single Neuron Is Sufficient to Bypass Safety Alignment in Large Language Models

SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training

ELF: Embedded Language Flows

PaperFit: Vision-in-the-Loop Typesetting Optimization for Scientific Documents

Rubric-based On-policy Distillation

CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Models

TMAS: Scaling Test-Time Compute via Multi-Agent Synergy

Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs

Qwen-Image-2.0 Technical Report

MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

Learning while Deploying: Fleet-Scale Reinforcement Learning for Generalist Robot Policies

Fast Byte Latent Transformer

AI Co-Mathematician: Accelerating Mathematicians with Agentic AI

HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents

Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers

LLMs Improving LLMs: Agentic Discovery for Test-Time Scaling