InternVL3.5-Flash Collection InternVL3.5-Flash is a fast variant of InternVL3.5 using semantic aware dynamic high-resolution strategy. • 9 items • Updated Oct 15, 2025 • 7
Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence Paper • 2604.18292 • Published 18 days ago • 83
Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders Paper • 2603.06569 • Published Mar 6 • 119
HISA: Efficient Hierarchical Indexing for Fine-Grained Sparse Attention Paper • 2603.28458 • Published Mar 30 • 44
Qwen3.5-text-only Collection Text-only versions of Qwen-3.5 without the vision encoders for a smaller memory and storage footprint. • 4 items • Updated 30 days ago • 15
Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels Paper • 2603.02573 • Published Mar 3 • 12
DualPath: Breaking the Storage Bandwidth Bottleneck in Agentic LLM Inference Paper • 2602.21548 • Published Feb 25 • 51
Micro-World Collection Action-controlled Interactive world model. • 3 items • Updated Feb 19 • 3
view article Article The Future of the Global Open-Source AI Ecosystem: From DeepSeek to AI+ Feb 3 • 53
SWE-Universe: Scale Real-World Verifiable Environments to Millions Paper • 2602.02361 • Published Feb 2 • 60
Falcon-H1-Tiny Collection A series of extremely small, yet powerful language models redefining capabilities at small scale • 19 items • Updated Mar 2 • 37
Cosmos-Reason2 Collection Cosmos Reason 2 is an open, customizable, reasoning vision language model (VLM) for physical AI and robotics • 8 items • Updated 8 days ago • 24
💧 LFM2 Collection LFM2 is a new generation of hybrid models, designed for on-device deployment. • 28 items • Updated 30 days ago • 154