LingBot VLA 4B Depth

LingBot-Video-Dense is a 1.3 billion parameter dense video generation model from the open source LingBot-Video family, built for embodied intelligence research. It supports text to image, text to video, and text and image to video generation, and is designed to connect video synthesis with physical world understanding rather than general entertainment use.

LingBot-Video-MoE is an open source Mixture-of-Experts video generation model with 30B total and 3B active parameters, built for embodied intelligence. It supports text-to-image, text-to-video, and image-to-video generation, plus a refiner stage for higher quality output, and is trained on web video combined with over 70,000 hours of embodied data.

LingBot VLA 2.0 is a 6B parameter Vision Language Action foundation model that converts camera images and language instructions into robot actions across arms, grippers, dexterous hands, waist, head, and mobile base signals for real world manipulation and mobile robot tasks.

SenseNova-U1-8B-MoT is SenseNova’s open native multimodal model for unified image understanding, reasoning, generation, and editing. It is built on the NEO-Unify architecture, uses an 8B dense MoT backbone, and supports text-to-image, image-to-text, image editing, and interleaved image-text generation in one model.

Grok 4.2 is referenced as a Grok 4-series update for xAI’s Grok assistant, focused on general-purpose help across reasoning, coding, and real-time info retrieval inside the Grok experience.

Multimodal embedding model that vectorizes interleaved text, images, and videos into a shared vector space for retrieval tasks. Uses a single transformer encoder for all modalities, eliminating the modality gap. Supports document screenshots, PDFs, slides, figures, tables, and video frames. Context window of 32,000 tokens with Matryoshka embeddings at 256, 512, 1024, and 2048 dimensions.