Spatial intelligence

Models 12

• AnnouncementGen 3 Mistral
  Robostral Navigate

  By Mistral AI

  Robostral Navigate is an 8B parameter robotics navigation model that uses a single RGB camera and natural language instructions to autonomously guide robots through complex environments. It predicts the next target location via image based pointing and falls back to local coordinate displacements when the target is out of view, achieving 76.6 percent success on R2R-CE validation unseen without depth sensors, LiDAR, or multiple cameras.

  🤖Autonomous mobile robots🧩Spatial intelligence👁️Computer vision assistance🤖Robotics
  NewMultimodal

  Released 13d ago
• Gen 3
  SANA WM (Bidirectional)

  By NVIDIA

  SANA-WM Bidirectional is Efficient-Large-Model’s open-source 2.6B image-to-video world model for 720p minute-scale video generation with 6-DoF camera control.

  🎥Videos🧩Spatial intelligence🎥Spatial image to video
  NewMultimodal

  Released 2mo ago
• Gen 3
  Perceptron Mk1

  By Perceptron AI

  Perceptron Mk1 is Perceptron’s closed-source vision-language model for video understanding, image reasoning, and embodied reasoning tasks.

  🎥Video analysis👀Video surveillance🧩Spatial intelligence🤖Robotics
  NewMultimodal

  Released 2mo ago
• Gen 3
  GENE 26.5

  By Genesis

  GENE-26.5 is Genesis AI’s first robotic foundation model system for general-purpose manipulation. It is built for long-horizon, contact-rich dexterity across real-world tasks such as cooking, lab automation, Rubik’s Cube solving, smoothie making, wire harnessing, and multi-object grasping, using one shared model, hardware platform, data strategy, and control stack.

  🤖Robotics🧩Spatial intelligence👁️Computer vision assistance🤖Robotic guidance
  NewMultimodal

  Released 2mo ago
• Gen 3
  MolmoAct 2

  By Ai2

  MolmoAct 2 is Ai2’s open robotics foundation model for real-world robot manipulation. It is an Action Reasoning Model that reasons about 3D structure before acting, supports bimanual manipulation out of the box, and is designed for zero-shot deployment and post-training adaptation across household, lab, and tabletop tasks.

  🤖Robotics🧩Spatial intelligence👁️Computer vision assistance
  NewMultimodal

  Released 2mo ago
• Gen 3
  MotuBrain

  By ShengShu Technology

  MotuBrain is ShengShu Technology’s world action model for embodied robotics. It is built as a universal brain for robots, unifying perception and action in one model so systems can understand environments, anticipate changes, and execute long-horizon tasks across homes, industry, and commercial spaces.

  🤖Robotics🤖Task automation🎲Decision making🧩Spatial intelligence🌎World creation
  NewMultimodal

  Released 2mo ago
• Gen 3
  Echo 2

  By SpAItial

  Echo-2 is SpAItial’s frontier 3D world-generation model for creating physically grounded, explorable environments from text or image inputs. Instead of generating a video, it produces a persistent 3D-consistent scene representation that can be rendered in real time on any device, targeting digital twins, robotics, architecture, gaming, and immersive 3D creation.

  🎨Scene creation🧩Spatial intelligence🤖Robotics🏛Architectural visualization🎥3D scenes
  NewMultimodal

  Released 2mo ago
• Gen 3
  HY Embodied 0.5 MoT 2B

  By Tencent

  HY-Embodied-0.5 is Tencent Hunyuan’s embodied foundation model suite for real-world agents and robots. It is built for spatial-temporal visual perception and embodied reasoning, including prediction, interaction, and planning, and serves as a "brain" for vision-language-action pipelines. The open release includes a Mixture-of-Transformers 2B edge model optimized for efficient deployment.

  🤖Robotics🧩Spatial intelligence👁️Computer vision assistance
  Multimodal

  Released 3mo ago
• Gen 1
  WildDet3D

  By Ai2

  WildDet3D is Ai2’s open model for monocular 3D object detection from a single RGB image. It predicts full 3D bounding boxes, including position, size, and orientation in metric coordinates, and supports flexible prompts such as text queries, point clicks, and 2D boxes. It is designed for open-world spatial understanding across varied cameras and can also use optional depth signals when available.

  🧩Spatial intelligence🔍Image recognition👁️Computer vision assistance
  3d

  Released 3mo ago
• Gen 3
  Isaac GR00T N1.7 3B

  By NVIDIA

  GR00T-N1.7-3B is NVIDIA’s open 3B vision-language-action foundation model for generalized humanoid robot reasoning and manipulation. It takes multimodal input including images, language, and proprioception, and outputs continuous robot actions, with support for post-training on specific robots or tasks using real or synthetic data.

  🤖Autonomous humanoids🧩Spatial intelligence🤖Robotics🤖Autonomous mobile robots
  Multimodal

  Released 4mo ago
• Gen 3 LingBot
  LingBot VLA 4B Depth

  By Robbyant

  LingBot VLA 4B Depth is a vision language action foundation model for robot manipulation. Pretrained on 20,000 hours of real world data from 9 dual arm robot configurations, it adds a depth distilled module for improved spatial perception, mapping camera images and language instructions directly to robot action sequences.

  🔍Image recognition🧩Spatial intelligence🤖Robotic guidance
  Multimodal

  Released 5mo ago
• Gen 4
  Depth Anything 3

  By ByteDance

  Depth Anything 3 is a spatial reconstruction model that predicts consistent geometry from arbitrary visual inputs, with or without known camera poses.

  🧩Spatial intelligence🌐2D to 3D image conversion👁️Computer vision assistance
  Image

  Released 8mo ago