TAPS: Task Aware Proposal Distributions for Speculative Sampling

LongTail Driving Scenarios with Reasoning Traces: The KITScenes LongTail Dataset

RealChart2Code: Advancing Chart-to-Code Generation with Real Data and Multi-Task Evaluation

Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills

PackForcing: Short Video Training Suffices for Long Video Sampling and Long Context Inference

ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytelling

Out of Sight but Not Out of Mind: Hybrid Memory for Dynamic Video World Models

BeSafe-Bench: Unveiling Behavioral Safety Risks of Situated Agents in Functional Environments

World Reasoning Arena

MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens

Voxtral TTS

RealRestorer: Towards Generalizable Real-World Image Restoration with Large-Scale Image Editing Models

Calibri: Enhancing Diffusion Transformers via Parameter-Efficient Calibration

Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale

PixelSmile: Toward Fine-Grained Facial Expression Editing

Claudini: Autoresearch Discovers State-of-the-Art Adversarial Attack Algorithms for LLMs

AutoHarness: Improving LLM Agents by Automatically Synthesizing a Code Harness

GameplayQA: A Benchmarking Framework for Decision-Dense POV-Synced Multi-Video Understanding of 3D Virtual Agents

Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs?

UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience

T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search

CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

Foveated Diffusion: Efficient Spatially Adaptive Image and Video Generation

Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos

From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agents

SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning

DA-Flow: Degradation-Aware Optical Flow Estimation with Diffusion Models

PEARL: Personalized Streaming Video Understanding Model

WildWorld: A Large-Scale Dataset for Dynamic World Modeling with Actions and Explicit State toward Generative ARPG

TAPS: Task Aware Proposal Distributions for Speculative Sampling

LongTail Driving Scenarios with Reasoning Traces: The KITScenes LongTail Dataset

RealChart2Code: Advancing Chart-to-Code Generation with Real Data and Multi-Task Evaluation

Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills

PackForcing: Short Video Training Suffices for Long Video Sampling and Long Context Inference

ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytelling

Out of Sight but Not Out of Mind: Hybrid Memory for Dynamic Video World Models

BeSafe-Bench: Unveiling Behavioral Safety Risks of Situated Agents in Functional Environments

World Reasoning Arena

MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens

Voxtral TTS

RealRestorer: Towards Generalizable Real-World Image Restoration with Large-Scale Image Editing Models

Calibri: Enhancing Diffusion Transformers via Parameter-Efficient Calibration

Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale

PixelSmile: Toward Fine-Grained Facial Expression Editing

Claudini: Autoresearch Discovers State-of-the-Art Adversarial Attack Algorithms for LLMs

AutoHarness: Improving LLM Agents by Automatically Synthesizing a Code Harness

GameplayQA: A Benchmarking Framework for Decision-Dense POV-Synced Multi-Video Understanding of 3D Virtual Agents

Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs?

UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience

T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search

CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

Foveated Diffusion: Efficient Spatially Adaptive Image and Video Generation

Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos

From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agents

SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning

DA-Flow: Degradation-Aware Optical Flow Estimation with Diffusion Models

PEARL: Personalized Streaming Video Understanding Model

WildWorld: A Large-Scale Dataset for Dynamic World Modeling with Actions and Explicit State toward Generative ARPG