Ant Group launches Ling-3.1-flash, a 560B-parameter model it plans to open-source

Ant Group's InclusionAI launched Ling-3.1-flash on 30 September 2026: 560 billion total parameters with about 25 billion active per token, aimed at agent tasks, search, office software and specialist applications. It is designed for a one-million-token context window, but the current two-week free trial is limited to 256,000 tokens; InclusionAI plans to enable the larger window and release the model as open source after the trial. It is also available on Vercel's AI Gateway.

What it means If the open release follows the trial, it would be one of the largest self-hostable mixture-of-experts models; watch whether it does.