Ant Ling today launched Ling-3.2-flash, a native multimodal model with 560B total parameters but only 25B activated at a time, featuring a 1M context window and following the MoE approach in which only part of the parameters are used during forward passes, preserving knowledge capacity while reducing per-step cost. The model architecture natively connects vision and language, handling images, text, and video alike, with four strengthened areas highlighted in the official release: complex visual understanding for deep reasoning over images and video; a multimodal long-horizon Agent that stays focused on goals across multi-step tasks and makes and executes decisions based on visual and textual input; 3D understanding to parse spatial relationships; and medical multimodal capabilities covering image interpretation and report generation. Ant's evaluation shows Ling-3.2-flash outperforms Claude Opus 5 (max) on three benchmarks: OmniDocBench1.5 scores 92.21 (vs. 90.56), ZEROBench(Sub) 43.71 (vs. 39.52), and Vision2Web 85.86 (vs. 80.20). In DesignArena's image-to-frontend arena, it entered the top ten with an Elo of 1238, sharing the stage with Meta, DeepSeek, and Anthropic. API pricing is $0.30 per million input tokens and $0.90 per million output tokens, rated by DesignArena