90 articles
GPT-4V(ision)
GPT-4V(ision) is the vision-enabled version of OpenAI's GPT-4, allowing the model to accept images alongside text; it reached ChatGPT users in September 2023 and the API that November.
GR (ByteDance robotics VLA family)
GR is a family of vision-language-action models for robot manipulation developed by ByteDance Seed, built on internet video pretraining and released as GR-1, GR-2, and GR-3 between 2023 and July 2025.
Grounded Language-Image Pre-training (GLIP)
Grounded Language-Image Pre-training (GLIP) is a vision-language model from Microsoft Research, introduced in December 2021, that unifies object detection and phrase grounding in one task.
GWM-1
GWM-1 is a family of generative world models announced by Runway on December 11, 2025, generating video frame by frame in real time with interactive control.
Helix (AI model)
Helix is a vision-language-action model developed by Figure AI and announced in February 2025 to control its humanoid robots end-to-end from camera pixels and natural-language commands.
Hunyuan3D
Hunyuan3D is a family of open 3D asset generation models from Tencent, first released in November 2024, that turns text or images into textured 3D meshes.
HunyuanWorld
HunyuanWorld is Tencent's open-source 3D world generation model family, first released in July 2025, converting text, images, or video into explorable 3D scenes importable into game engines.
InternVideo
InternVideo is a family of open video foundation models for video understanding, developed by the General Vision Group at Shanghai AI Laboratory since December 2022.
InternVL
InternVL is a family of open-weight vision-language models, processing images and text together, developed by OpenGVLab at Shanghai AI Laboratory and first released in December 2023.
Janus-Pro
Janus-Pro is a unified multimodal understanding-and-generation model released by DeepSeek on January 27, 2025, in 1B and 7B sizes, generating images from text within a single autoregressive model.
Kimi-VL
Kimi-VL is an open-weight Mixture-of-Experts vision-language model released by Moonshot AI in April 2025, notable for activating only 2.8B of its 16B parameters while offering multimodal reasoning, long-context understanding, and agent capabilities.
Kosmos
Kosmos is a family of multimodal large language models released by Microsoft Research in 2023, with Kosmos-2 adding pixel-level grounding of text to image regions.
LLaVA
LLaVA (Large Language and Vision Assistant) is an open-source family of vision-language models, first released in April 2023, that connects a CLIP vision encoder to a language model.
Magma (multimodal agent foundation model)
Magma is an 8-billion-parameter multimodal foundation model for AI agents released by Microsoft Research in February 2025, combining vision-language understanding with action-taking in digital interfaces and robotic manipulation.
Marble
Marble is a multimodal world model by World Labs (San Francisco). Founded by Fei-Fei Li, it is the first commercial AI to generate navigable 3D environments from diverse data.
Matrix-Game
Matrix-Game is a series of open interactive world foundation models from Skywork AI that generate playable game environments from a reference image and user actions, first released in 2025.
Matrix-Game 2.0
Matrix-Game 2.0 is an open-source interactive world model released by Skywork AI in 2025, streaming navigable video at 25 frames per second under an MIT license.
MedSAM
MedSAM is a promptable AI foundation model for medical image segmentation, fine-tuned from Meta's Segment Anything Model on over 1.5 million medical image-mask pairs and introduced in April 2023.
MiniCPM-V
MiniCPM-V is a family of open-weight multimodal large language models developed by OpenBMB, first released in February 2024 and designed to run image and video understanding on consumer devices like smartphones.
MiniGPT-4
MiniGPT-4 is an open-source vision-language model developed at KAUST and introduced in 2023, aligning a frozen vision encoder with the Vicuna language model through one projection layer.
Molmo
Molmo is a family of open-weight vision-language models released by the Allen Institute for AI in September 2024, distinguished by a pointing capability and fully open training data.
Motubrain
Motubrain is a World Action Model for robot control released on 29 April 2026 by Chinese AI company Shengshu Technology, jointly modeling video prediction and robot action generation.
mPLUG-Owl
mPLUG-Owl is a series of open-source multimodal large language models from Alibaba's DAMO Academy, first released in April 2023, with three generations through 2024 covering images and video.
Neural radiance fields
Neural radiance fields (NeRF) is a technique for synthesizing novel views of a scene from ordinary photographs, introduced by Ben Mildenhall and colleagues in March 2020.
Nomic Embed
Nomic Embed is a family of fully open embedding models, text, vision, and multimodal, released by Nomic AI from February 2024 with weights, code, and data all public.
NVIDIA Cosmos
NVIDIA Cosmos is a family of open world foundation models from NVIDIA for physical AI, predicting and simulating world states to train robots and autonomous vehicles.
NVIDIA Isaac GR00T
NVIDIA Isaac GR00T, also known as Project GR00T, is an open family of vision-language-action foundation models for humanoid robots, first released by NVIDIA in March 2025.
Oasis
Oasis is a real-time AI world model released by Decart and Etched in October 2024, generating a playable, Minecraft-style world frame-by-frame with no game engine.
Octo (robot foundation model)
Octo is an open-source generalist robot policy that maps camera images and task specifications to robot actions, pretrained on 800,000 demonstrations and released in May 2024.
OpenAI text-embedding-3
text-embedding-3 is a family of OpenAI text embedding models released in January 2024, in small and large versions that convert text into vectors for search, clustering, and retrieval-augmented generation.