AntGroup releases Ling 3.0 Flash MoE model

Today, we’re releasing Ling-3.0-flash—a hybrid-reasoning MoE model built for production-scale agents. 124B parameters. Just 5.1B active per token. With 1/8 of the total and 1/12 of the active parameters, it matches or beats our 1T flagship model on most benchmarks shown.

• Ling-3.0 starts with native hybrid-linear attention: KDA and MLA layers stacked 5:1. KDA gives fine-grained control over long-range memory, while 1/64 expert activation makes MoE compute more efficient. It supports 256K context natively and can scale to 1M.Ling-3.0-flash is now live on OpenRouter—and free to use through August 3, 2026. Try it in your coding, search, research, and tool-use workflows. Then show us what you build:Demo 1 — From one prompt to a 3D world. Using Blender MCP, Ling-3.0-flash wrote Python, built a city with elevated roads, skyscrapers, and materials, set the camera path, and rendered an aerial video—showing spatial reasoning and long-horizon tool use.Demo 2 — An autonomous research team. Ling-3.0-flash coordinated Scientist, Data Analyst, CrossValidator, Archivist, and Writer agents to form hypotheses, search literature, challenge evidence, resolve disagreements, and produce a paper plus slide deck.Demo 3 — Office work, end to end. Through Office MCP, Ling-3.0-flash read and formatted a Word proposal, generated its table of contents, checked Excel financials with SUMIF and pivot tables, and exported polished documents—without manual cleanup.Demo 4 — Design systems from code alone. Without external images, Ling-3.0-flash generated webpages across Bauhaus, Bohemian, acid design, and more—using CSS gradients, SVG paths, typography, and layout to preserve each visual language.Demo 5 — A multimedia app in the browser. Ling-3.0-flash built a minimalist piano and synthesizer with multiple waveforms, envelope controls, and real-time Canvas audio visualization—combining frontend engineering, interaction design, and Web Audio.Demo 6 — One context, three channels. Given a real-estate sales email, Ling-3.0-flash extracted the key facts and turned them into platform-native marketing copy for Instagram, X, and LinkedIn—adapting tone and format without losing the source context.Demo 7 — Scheduling that closes the loop. Connected through OpenClaw, Ling-3.0-flash checked the calendar, identified available course slots, replied with options, sent a confirmation, and updated the calendar—turning a message into a completed workflow.Reliable inference for the community. Ling-3.0-flash’s launch on @OpenRouteris supported by@novita_labs, bringing its 256K context and agentic capabilities to builders at scale. Free access runs through Aug 3. Open-source release coming soon—stay tuned. 🤝 - any open weights? I noticed you forgot the hugging face link...