Video-as-Answer: Predict and Generate Next Video Event with Joint-GRPO

This repository summarizes recent advances in the VLA + RL paradigm and provides a taxonomic classification of relevant works.

RLinf is a flexible and scalable open-source infrastructure designed for post-training foundation models (LLMs, VLMs, VLAs) via reinforcement learning.

Official implementation of "OmniX: From Unified Panoramic Generation and Perception to Graphics-Ready 3D Scenes".

LongLive: Real-time Interactive Long Video Generation

Official Repository for MolmoAct

This is the official repository for the paper "FLUX-Reason-6M & PRISM-Bench: A Million-Scale Text-to-Image Reasoning Dataset and Comprehensive Benchmark"

AudioStory: Generating Long-Form Narrative Audio with Large Language Models

T2I-ReasonBench: Benchmarking Reasoning-Informed Text-to-Image Generation

Structured Video Comprehension of Real-World Shorts

[NeurIPS 2025] OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene Understanding

Official implementation of the paper: "StreamVLN: Streaming Vision-and-Language Navigation via SlowFast Context Modeling"

[ICCV 2025] Official implementation of the paper "DreamCube: 3D Panorama Generation via Multi-plane Synchronization".

TokLIP: Marry Visual Tokens to CLIP for Multimodal Comprehension and Generation

Video-Holmes: Can MLLM Think Like Holmes for Complex Video Reasoning?

[ICCV 2025] Official repo for "GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation"

HoloPart: Generative 3D Part Amodal Segmentation

[ICCV 2025] AnimeGamer: Infinite Anime Life Simulation with Next Game State Prediction

[ICCV2025] TokenBridge: Bridging Continuous and Discrete Tokens for Autoregressive Visual Generation. https://yuqingwang1029.github.io/TokenBridge

[ICCV 2025] GameFactory: Creating New Games with Generative Interactive Videos

[CVPR2025 Highlight] PAR: Parallelized Autoregressive Visual Generation. https://yuqingwang1029.github.io/PAR-project

[Neurips 2023 & TPAMI] T2I-CompBench (++) for Compositional Text-to-image Generation Evaluation

[CVPR 2025] T2V-CompBench: A Comprehensive Benchmark for Compositional Text-to-video Generation

[CVPR 2024] DreamComposer: Controllable 3D Object Generation via Multi-View Conditions

SAMPart3D: Segment Any Part in 3D Objects