Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters Paper • 2602.10604 • Published 23 days ago • 187
DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal Reasoning Paper • 2602.16742 • Published 16 days ago • 11
Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text Paper • 2601.22975 • Published Jan 30 • 109