Ant's Ling-3.2-flash debuts: 560B/25B MoE, 1M context, 85.63 on AndroidWorld
Beginning-Wealth-480 · reddit · 2026-11-09
Ant Ling released Ling-3.2-flash: a 560B-total/25B-active model with 1M token context, native text/image/video input, and tool use via OpenAI- and Anthropic-compatible APIs. It scores 85.63 on AndroidWorld and 67.77 on BlenderBench, with demos of driving Blender via mouse and keyboard to build a furnished room.
More from Models
- Open-weights Inkling Small 2 impresses: researcher says he prefers it over DeepSeek V4.1 Flash — gnukeith · 2026-11-13
- Manual Review Exposes How LLMs Cheat Benchmarks: Sandbox Escapes and Copying Peers — teortaxesTex · 2026-11-13
- Models invent their own filler tokens during RL training — code_star · 2026-11-13
- Google Search reportedly gets massive Gemini 4 Hydrogen upgrade with in-search simulations — AI_Andrew · 2026-11-13
- Google ships Gemini 4 Hydrogen: beats its bigger Argon model on intelligence index, 3x coding gains — 量子位 · 2026-11-13
- Chinese medical AI beats 250 top doctors 88-12 on complex lung cancer cases in new study — 量子位 · 2026-11-13