A general vision-language model can serve as a multimodal reranker via single-pass choice questions, matching or exceeding ...
Matrix-Game 4.0 separates game logic from visual synthesis via an executable engine and autoregressive video renderer, ...
SolarRenderer turns any engine's untextured structure pass into a finished, styled video, guided by a reference first frame.
MMEnterprise-Bench evaluates this challenge with 1,613 questions across 30 simulated collaboration projects, grounded in ...
🤖 GroundingPI: What foundation model does embodied AI need? 4B parameters. SOTA overall across 34 grounding benchmarks, surpassing GPT-6 Astra! By training precise perception as a native capability, ...
Describe Anything Model 3B (DAM-3B) takes inputs of user-specified regions in the form of points/boxes/scribbles/masks within images, and generates detailed localized ...
Author here. WebFovea is a vision-based web agent that placed 2nd in the WebRetriever Challenge 2026 (57.0/100). It works on live websites through their own UI: screenshots in, clicks and keystrokes ...
What changes when memory, exact execution, and contextual computation can be developed and updated separately? This research grew out of experiments showing that useful language modeling can emerge ...
Add files using upload-large-folder tool about 1 year ago meta Add files using upload-large-folder tool about 1 year ago videos Add files using upload-large-folder tool about 1 year ago ...
SpecMRL improves compressed embeddings by reorganizing features via a learned spectral basis that prioritizes informative ...
AdaExploreBench evaluates whether LLM agents can actively discover hidden environmental rules and adapt to unannounced ...
MIST separates scene setup from change execution to isolate memory failures in video world models, and M3-Agent improves ...
Results that may be inaccessible to you are currently showing.
Hide inaccessible results