Say what?
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, generative and decision models.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
me and my d making decisions
Today we release 𝚍𝟷-𝚞𝚕𝚝𝚛𝚊, our most powerful decision model.
> built for hard decision problems.
> d1-ultra is multimodal
> supports 44 languages
> #1 on the @ArtificialAnlys Decision Index (65)
> leads all eight categories of our 112-evaluation suite
1/
i could smell the cooking miles away
Ty @liquidai
nitter.cf/liquidai/status/211005…
I cant wait for the the next one 👑🐐
Announcing d1, our first decision model.
It's the first model to outperform Jev on @huggingface's Decision Index.
> wins on multilingual evals
> more robust against prompt injection
> handles longer inputs more effectively
> built for fast, structured decision-making in software environments
> Liquid API: console.liquid.ai
> Available on OpenRouter soon
Show a person something their mind cannot forget, and they will begin to believe in it. Because once they have seen it, denying it would mean denying their own intellect. That is one of the greatest traps.
Kids are going to be born directly into the AI genjutsu, and I’m genuinely not sure how that plays out.
slowly, @liquidai will power more machines than @AnthropicAI
not if, when.
their 450M-param model already runs wildfire risk detection on satellite imagery, on-device, sub-250ms, no cloud round trip needed
small models doing real jobs while everyone argues about agent skills
i've been jumping for a while, you should too.
waitlist.jumper.xyz/r/JMP-92…
Nothing humbles you like watching the market discover ideas you were talking about a year ago.
Encoders. Specialized SLMs outperforming general LLMs.
Apparently being consistently one year ahead of trends is just my favorite form of self-punishment.
bioxbt retweeted
Find benchmarking protocol details here: arxiv.org/abs/2608.18940, the new GPT-6 Astra results will be soon released at dddbench.insilico.com 2/2
