Liquid AI

We’re on our way, Tokyo! Apply to join our 2-day hackathon co-hosted with @wayequity and @AMD. Engineers, founders, and mentors across the Liquid AI and WAY ecosystems will gather to ship real-world applications to accelerate Japanese industry, powered by our Liquid Foundation Models (LFMs). Selected participants, in teams of 1-3, will create applications/workflows to address real-world problems only possible with LFMs. Top projects will be awarded: / Gold Prize - $3K USD / Silver Prize - $2K USD / +Internship offers, community recognition, and more.

More conversations with @mlech26l and the team next week! Follow us here for every interview.

Building a model is just the start. Post-training makes it useful. Our CTO Mathias Lechner (@mlech26l) sits down for a conversation with Maxime Labonne (@maximelabonne), our head of post-training, on the pipeline that takes a base model from autocomplete to something that can reason and follow instructions.

Listen to the full episode: podcasts.apple.com/us/podcast/why…

Well said. AI shouldn't just live in a data center, but it should run on the device you're already using: Whether it's a phone, a car, an appliance, or another everyday device in the real world. That’s what we’re working on at Liquid AI. We appreciate the spotlight on our COO Jeffrey Li (@jeffr3yli), who joined the AI-Curious Podcast (@AiCuriousHQ). Thank you, @jeffwilser!

Over the next few weeks, @mlech26l will be in conversation with our researchers and engineers on the work they're shipping, from training and inference infrastructure to on-device deployment, multimodal models, biology, and more. Follow along here for every interview!

AI demand is growing exponentially. The energy to power it isn't. Kicking off a new series where our CTO Mathias Lechner (@mlech26l) sits down with the team to talk about what they're building. First up: Our CEO Ramin Hasani (@ramin_m_h) on the demand-energy gap, and what it takes to close it.

I might have found my new local LLM, and I don't see a lot of people talking about it, if you have more data and context about this model please let me know: I tested LFM2 from @liquidai. that's a hybrid SSM + conv + MoE, only 10 of 40 layers use attention, the rest are state-space layers. 2.3B active params out of 24B total. my results: > 52.2 tok/s decode at 32K context > ncmoe 22 sweet spot (18 of 38 expert layers on GPU) > 13.4 GB Q4_K_M - lightest of all three-way head-to-head (same rig, same method, same quant): > LFM2: 52.2 tok/s - new champion > Qwen3.6 35B A3B: 35.4 tok/s > Gemma 4 26B A4B: 29.3 tok/s I then ran 12 stress tests beyond speed: JSON compliance, reasoning, coding, long context, multi-turn tracking, repetition. quality is solid for technical work, but it fails on creative diversity at temp=0 and cross-entity reasoning. but hey, on 8GB VRAM, it works pretty damn well! SSM layers need no KV cache, so context scaling is nearly free. 65K at 37.6 tok/s while Qwen drops to 17.4. data + methodology on HF: huggingface.co/datasets/witch…

Apply to join: luma.com/7fjlam5k Start building with LEAP: leap.liquid.ai

東京ハッカソン２.０！ @WAYequity、@AMDとの共催による2日間のハッカソンへの参加応募を受付中です。 Liquid AIとWAYのエコシステムに関わるエンジニア、創業者、メンターが集まり、Liquid Foundation Models（LFMs）を活用して、日本の産業を加速させる実用的なアプリケーションの開発に取り組みます。 選抜された参加者は、1〜3名のチームで、LFMsだからこそ実現できる現実世界の課題解決に向けたアプリケーションやワークフローを作成します。 優秀プロジェクトには以下が授与されます。 / ゴールド賞：3,000米ドル / シルバー賞：2,000米ドル / インターンシップのオファー、コミュニティでの表彰、その他特典もあります。

We’re on our way, Tokyo! Apply to join our 2-day hackathon co-hosted with @wayequity and @AMD. Engineers, founders, and mentors across the Liquid AI and WAY ecosystems will gather to ship real-world applications to accelerate Japanese industry, powered by our Liquid Foundation Models (LFMs). Selected participants, in teams of 1-3, will create applications/workflows to address real-world problems only possible with LFMs. Top projects will be awarded: / Gold Prize - $3K USD / Silver Prize - $2K USD / +Internship offers, community recognition, and more.

Advice for AI engineers 💡 If you're building voice agents, stop wiring up 3 separate models, for audio-to-text, text-to-audio, or text-to-text. 𝗹𝗶𝗾𝘂𝗶𝗱-𝗮𝘂𝗱𝗶𝗼 is the open-source repo for SOTA speech-to-speech LFM models. End 2 end. Now with fine-tuning support. More examples coming. Bookmark this ↓ github.com/Liquid4All/liq…

AI infra has a memory problem and we have a Liquid solution for it! open.substack.com/pub/raminmh/p/…

Everything I Learned Training Frontier Small Models @maximelabonne After a lot of hype around frontier model training, Maxime gives the practical version: what actually happens when you try to train smaller models that still matter. From data quality and synthetic data to evals, distillation, and where small models still punch above their weight, this is a concise field report from someone doing the work at @liquidai. If you're building with open models, post training, or trying to understand where the next wave of useful local-ish capability comes from, this one is worth your time. youtube.com/watch?v=fLUtUk…

on device AI communicating with us from outer space!

LFM2:3B in space, on Cluster Gate2: ✨ “This image is a highly detailed, close-up view of Earth as seen from space, likely captured by a satellite or space telescope. The Earth is depicted as a large, circular sphere with a predominantly blue hue, indicating the vast oceans that cover most of its surface. The blue is interspersed with swirling white clouds, which are particularly prominent over the landmasses, suggesting the presence of weather systems and atmospheric activity. The overall composition of the image highlights the beauty and complexity of our planet, showcasing the dynamic interplay between the oceans, atmosphere, and landmasses." Congratulations to @DPhiSpace for this incredible milestone! 🌎

Some say Gated DeltaNet > Mamba-2. Others say Mamba-2 > Gated DeltaNet. But what if Gated DeltaNet = Mamba-2? 👀 Well maybe not exactly — but with least-squares preconditioning, we show that they reduce to the same recurrence! We use this lens to design PDN, PGDN, and PKDA: preconditioned delta-style recurrences that outperform their unpreconditioned counterparts at scale 📈 📄 Paper: arxiv.org/abs/2604.21100 w/ @loo_noel @liquidai 💻 Code: github.com/ntumm120/preco…

@ramin_m_h 👀

another big thing coming soon! 👣🪧

Read all about our partnership with @MercedesBenz, announced today: media.mbusa.com/releases/merce…

Start your engines. The news about our partnership with @MercedesBenz is lighting up Times Square today. Thank you to Brian Baumann (@bjbaumann2014) and the @NYSE for the support!

and we got a shoutout on our partnership with Mercedes Benz today at the Times Square from @NYSE! 🎉🏁 Read the announcement: media.mbusa.com/releases/merce…