Ant Ling releases Ling-3.2-flash, a 560B-parameter native multimodal model

AIDesignArena says Ant Ling's Ling-3.2-flash ranks #10 on its Image to Frontend Arena with an Elo of 1238. The model costs $0.30 per 1M input tokens and $0.90 per 1M output tokens, which DesignArena calls the cheapest for its performance level. It wins most often on Portfolio and Corporate builds, at about 63% and 60% of tournaments respectively.

AIAnt Ling demonstrates Ling-3.2-flash coordinating specialist agents with shared context to turn Q2 financial data into an analysis report, a six-slide board deck and a highlights video. The demo also shows the agents refining those deliverables in response to user feedback.

AIAnt Ling introduces Ling-3.2-flash, a native multimodal model with 560B total and 25B active parameters. The company says it outscores Opus 5 (max) in release evaluations, including 92.21 versus 90.56 on OmniDocBench1.5, 43.71 versus 39.52 on ZEROBench (Sub), and 85.86 versus 80.20 on Vision2Web. It targets complex visual understanding, long-horizon multimodal agents, 3D understanding and interaction, and medical intelligence.