Skip to content
Neural Daily

ModelsNews

Ant Ling-3.2-flash: 560B MoE, 1M context, multimodal

🇭🇺 Olvasd el magyarul →

Ant Group released Ling-3.2-flash, a native multimodal MoE with 560B total / 25B active parameters, a 1M-token window and text, image and video input. Chat plus OpenAI/Anthropic-compatible APIs are live, free for two weeks. Weights stay closed.

· 1 min read

Fehér humanoid robot ujjal mutat egy falra, AI-illusztráció
Image: Pexels / Tara Winstead · Pexels · Source

What Ling-3.2-flash adds

Per the Ant Ling docs, Ling-3.2-flash is the series' latest native multimodal model: 560B total parameters, about 25B active per step, and a 1M-token context. It takes text, image and video without an external vision module.

Four focus areas: complex visual understanding, multimodal long-horizon agents, 3D understanding and interaction, and medical multimodal work. TechNode says it launched November 5 with chat plus OpenAI/Anthropic-compatible APIs, free for the first two weeks.

What not to expect

Unlike the earlier MIT-licensed Ling-3.1-flash-1019, Ling-3.2-flash weights are not open; access stays on Ant's platform. The changelog dates the 3.2 series to November 2026.

Key takeaways

  • 560B / 25B MoE, 1M context, native image+video.
  • Focus: visual reasoning, long-horizon agents, 3D, medical MM.
  • API+chat live; two weeks free; weights closed.

    By Mikola Máté Patrik

    Produced with the help of AI, reviewed by an editor. How we make our articles

    Reviewed and edited by the Neural Daily desk. About

    Related articles