Ant Ling-3.1-flash — open weights under MIT licence
Ant Group's inclusionAI published Ling-3.1-flash weights on Hugging Face under an MIT licence: a 560B-total/25B-active mixture-of-experts model with 1M-token context aimed at office work and long-horizon agents, following the end of its two-week free API trial on Oct 14.
Why it matters
A 560B MoE under the MIT licence with 1M context and vendor-reported ~210 tok/s throughput — one of the most capable openly licensed models aimed squarely at real office and agent workloads rather than chat demos.
What you could build with it
A startup building an AI paralegal for contract review could self-host Ling-3.1-flash to process hundred-page documents in a single context window under a permissive MIT licence, avoiding per-token API costs on long-document workloads.
Does it hold up?
Too early to judge: weights dropped today under MIT. Ant's launch came with vendor-reported benchmarks only and no third-party evals yet, so real-world quality is unverified.
Built with Ant Ling-3.1-flash
- Ling-3.1-flash | The Deploy Logarticle · Independent deployment-tracker entry: weights on Hugging Face, ModelScope, and OpenRouter; model published Oct 13, added to the log Oct 14.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74StepFun Step-5-Preview — open BF16 weights releasedStepFun published the full BF16 weights for its flagship Step-5-Preview agentic model (600B-parameter sparse MoE, 27B…models · JEV 0.73ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71JetBrains Mellum 2.1: 12B MoE coding model trained for agentsJetBrains released Mellum 2.1, a 12B MoE model (2.5B active parameters, 128K context) under Apache 2.0, trained mainly…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.