baaivision

EVA Series: Visual Representation Fantasies from BAAI

Painter & SegGPT Series: Vision Foundation Models from BAAI

Next-Token Prediction is All You Need

Emu Series: Generative Multimodal Models from BAAI

Native Multimodal Models are World Learners

[CVPR'25 Highlight] You See it, You Got it: Learning 3D Creation on Pose-Free Videos at Scale

[ICLR'24 Spotlight] Uni3D: 3D Visual Representation from BAAI

[ICLR 2025] Autoregressive Video Generation without Vector Quantization

[ECCV 2024] Tokenize Anything via Prompting

GeoDream: Disentangling 2D and Geometric Priors for High-Fidelity and Consistent 3D Generation

[ICLR 2025 Spotlight] An open-sourced LLM judge for evaluating LLM-generated answers.

EVE Series: Encoder-Free Vision-Language Models from BAAI

Zero-Shot Video Editing Using Off-The-Shelf Image Diffusion Models

[ICLR 2026] Unified Vision-Language-Action Model

[ICLR 2025] Diffusion Feedback Helps CLIP See Better

[CVPR 2024] CapsFusion: Rethinking Image-Text Data at Scale

DenseFusion-1M: Merging Vision Experts for Comprehensive Multimodal Perception

[ICLR 2026] 🐻 Uniform Discrete Diffusion with Metric Path for Video Generation

An in-context conditioning version of MUSE with pre-trained checkpoints.

MTVCraft: An Open Veo3-style Audio-Video Generation Demo