{"slug": "liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment", "title": "Liquid AI unveils LFM3 family of customizable AI models for local deployment", "summary": "Liquid AI launched the LFM3 family on October 27, 2026: six open-weight multimodal models of roughly 2 billion parameters each, built to run and be fine-tuned on a single GPU. All six share the LFM3-1.7B-Base backbone and a tri-hybrid architecture combining gated short convolutions, sliding-window attention and global attention, pre-trained on 22 trillion tokens; the text model LFM3-1.7B supports a 256K-token context window, and Liquid AI claims LFM3-VL-2B matches models more than six times its size on vision tasks. Liquid AI, founded in 2023, has raised approximately $287.5 million, including a $250 million Series A in December 2024, and is targeting domain-specific edge uses such as robotics and defect detection.", "body_md": "Photo: Merlin Lightpainting / Pexels\n\n# Liquid AI unveils LFM3 family of customizable AI models for local deployment\n\nSix open-weight multimodal models of roughly 2 billion parameters each are built to run and fine-tune on a single GPU\n\nLiquid AI is betting that the next big thing in artificial intelligence is actually pretty small. On October 27, 2026, the company launched its LFM3 family: six open-weight multimodal models built to run and be fine-tuned on local hardware.\n\nEach model carries approximately 2 billion parameters. In an industry where model sizes often get announced like stadium attendance figures, that is a deliberately modest number.\n\nThe pitch is simple. You should not need a data center to get useful AI, and you should be able to customize it on a single GPU.\n\n## Six models, one backbone\n\nAll six models share a common foundation called LFM3-1.7B-Base. Think of it as one chassis with different bodies bolted on top, each tuned for a different kind of input.\n\nThe lineup spans text, vision-language, audio, decision-making, and an “omni” model. That omni variant handles text, images, video, and audio together, which makes it the Swiss Army knife of the bunch.\n\nThe text model, LFM3-1.7B, supports a context window of 256K tokens. A context window is how much material a model can hold in its head at once, so a larger one means it can work through long documents without losing the thread.\n\nThe vision-language model, LFM3-VL-2B, is where Liquid AI is making its boldest claim. According to the research findings, it has shown performance matching models more than six times its size on vision tasks.\n\n### AI, tech, and the markets they move—in one daily briefing.\n\nDaily. Free. Join 34,000+ readers across crypto, finance, and policy.\n\nBecause the models are open-weight, developers can download them and adapt them to their own needs. Liquid AI is highlighting domain-specific uses such as robotics and defect detection, the sort of jobs where a model sits close to the action rather than in a distant cloud.\n\n## The tri-hybrid architecture\n\nThe technical centerpiece is what Liquid AI calls a tri-hybrid architecture. It combines three components: gated short convolutions, sliding-window attention, and global attention.\n\nGated short convolutions handle the nearby, local patterns efficiently. Sliding-window attention looks at a moving chunk of recent context, while global attention lets the model connect ideas across the entire input.\n\nThe architecture was pre-trained on 22 trillion tokens. The result, according to the research findings, is a high performance-per-parameter ratio compared with larger models.\n\n## Who is Liquid AI\n\nLiquid AI was founded in 2023 and has raised approximately $287.5 million since then. The largest piece of that came from a $250 million Series A round in December 2024.\n\nEdge hardware refers to devices that process data where it is created, such as robots, cameras, and factory equipment. Running AI on those devices directly can cut latency and keep sensitive data from ever leaving the building.\n\n## What this means\n\nFor developers and enterprises, the most practical detail is the single-GPU fine-tuning. A manufacturer that wants a model to spot defects on its own production line does not necessarily need a general-purpose giant. It needs something small, cheap to adapt, and good at one specific job. That is precisely the niche LFM3 is targeting.\n\nThe open-weight approach also matters. Teams can run the models on their own hardware, inspect them, and tailor them without routing every request through a third-party service.\n\nThere are caveats worth watching. Benchmark claims, including the six-times-larger comparison for LFM3-VL-2B, will be tested by independent developers once the models are in wider use, and real deployments often look messier than lab results.\n\n**Disclosure:** This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our\n\n[Editorial Policy](https://cryptobriefing.com/editorial-policy/).", "url": "https://wpnews.pro/news/liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment", "canonical_source": "https://cryptobriefing.com/liquid-ai-lfm3-local-ai-models/", "published_at": "2026-10-27 17:44:39+00:00", "updated_at": "2026-10-27 18:15:13.229531+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-startups", "ai-infrastructure", "ai-tools"], "entities": ["Liquid AI", "LFM3", "LFM3-1.7B-Base", "LFM3-1.7B", "LFM3-VL-2B", "Diego Almada Lopez"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment", "markdown": "https://wpnews.pro/news/liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment.md", "text": "https://wpnews.pro/news/liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment.txt", "jsonld": "https://wpnews.pro/news/liquid-ai-unveils-lfm3-family-of-customizable-ai-models-for-local-deployment.jsonld"}}