Ling 3.2 Flash: Ant Group's 560B native multimodal MoE, now free on OpenRouter
Ant Group's inclusionAI launched Ling-3.2-flash, a 560B-parameter MoE (25B active per token) that reasons natively over text, image and video with a 1M-token context window; announced 5 Nov 2026, it landed on OpenRouter as a free model on 13 Nov.
Why it matters
First Ling model that treats image and video as first-class reasoning inputs rather than bolted-on vision, at 560B scale with a 1M context, free via API for its first two weeks.
What you could build with it
A small team could build a document-intelligence service that feeds whole scanned archives, hundreds of pages of invoices and contracts, into Ling-3.2-flash's 1M-token window in a single pass, extracting structured data without chunking, and charge per document processed.
Does it hold up?
Too early to judge - the API is live and free until about 19 Nov, but the weights are not released and no independent benchmark has landed; vendor claims of beating Opus 5 on three evals are unverified.
Built with Ling 3.2 Flash
- Ling-3.2-flash: Free 560B Multimodal Model, API and Specsfreakyjolly.com · Guide to the model, its free two-week API window and OpenAI-compatible endpoints.
- Ling 3.2 Flash: 560B Multimodal, a 30-Second Video Caporcarouter.ai · Independent look at the spec sheet and the limits it hides: a 30-second video cap, 32-frame sampling, and no published price yet.
- Ling-3.2-flash - 1M context, multimodalllmreference.com · Spec sheet: 560B/25B MoE, 1M context, proprietary licence, weights not released.
Learn more
First spotted on article: source.
More AI models
VoiceStudio: open-source, fully-local ElevenLabs alternative for voice cloning, dubbing, and audiobooksdebpalash/VoiceStudio is a desktop app that runs voice cloning, voice design, video dubbing, dictation, transcription…models · JEV 0.75Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Google AI Edge Gallery: offline speech recognition and synthesis on-deviceGoogle added fully offline speech recognition and text-to-speech to AI Edge Gallery, with ML Kit APIs on Android and…models · JEV 0.73Soup: fine-tune an 8B LLM on a 4 GB laptop GPU from one YAML fileAn open-source CLI that fine-tunes LLMs from a single YAML file, with layer streaming that trains an 8B model on a 4…models · JEV 0.73Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.