Vishnu Samavedula

Architect @liquidai | ex-Stripe

[Vishnu Samavedula @svrvishnu]

Astra maxing :)

[Leonie @helloiamleonie]

we published a blog on hugging face at possibly the worst time yesterday lol

congrats to the HF team on the big news! 💚

here's a fine-tuning tutorial showing how to
• fine-tune a tiny LFM2.5-350M model
• in 100 GRPO steps using TRL
• for better structured outputs

blog: https://t.co/0ijRPNQdzH
colab: https://t.co/QAAPY5a6Vn

[Vishnu Samavedula @svrvishnu]

@vivmarquez 😁

[Ramin @ramin_m_h]

about a year ago, we released the first instances of Liquid nanos. these are products we sell to enterprises: tiny models + the customization platform matching the quality of frontier models on specialized use-cases. It is easy to recursively customize a model to 90% production quality. going from 90 to 100 that satisfies enterprise reliability, and QA is the art. It is very hard, but joyfully doable

Instances of these Liquid nanos are today in full production at @Shopify at scale!

https://t.co/LWj3uSS6Li

[Julien Chaumond @julien_c]

open source ai must win

[Vishnu Samavedula @svrvishnu]

@cdngdev No ducks were harmed in this experiment 🫣😆

[Vishnu Samavedula @svrvishnu]

@auchenberg Did u get a chance to try out the micro keyboard ?

[0xSero @0xSero]

Top Models For your Hardware 2026

-- 8GB --

- lfm-2.6B: very solid, trained on tons of data, obviously limited by good place to start

Use this for snippet tool calls, or to play around and build infra https://t.co/nCsHG1Zmrx

-- 16GB --

I am a big fan of Ornith models, they do well at tool calling and they certainly push models farther

Gemma is a more well rounded model for chat, and vision IMO
https://t.co/tc6IBTrbc3
https://t.co/n9HS0emaKR

-- 24GB up to 96GB --

Qwen3.8-27B is IMO the first model here you can code with, it's really strong. Use the exl3 versions.
https://t.co/n5avYaHKD2

-- 96GB up to 196GB --

Qwen3.8-Flash-Next is where it's at. Phenomenal model, very fast even on slower hardware, kv cache is smaller and part of the model is basic enough to be offloaded to ram https://t.co/ZHF8HdNBG7

-- 196GB up to 384GB --

GLM-5.3-Flash is frontier at home, built to run on 10,000$ of hardware. It's really a gift. It is natively multimodal, image/video/ audio?

It hold up over 1 million tokens in context, and is hybrid attention, it holds on higher concurrency. Very good for 3D, coding, hacking, art.

https://t.co/hf9ZjwGtca

-- 384GB up to 512GB --

Sacrifice speed for intelligence, GLM-5.3 is tied for #1 Open Weight model, and is the best for pure coding and systems work.

https://t.co/ujKpGbAwGN

I told you August was going to be awesome, now we go to slower months.

-------------

It's not Local AI we have to worry about anymore.

[Vishnu Samavedula @svrvishnu]

🙃 i did a thing !!...now waiting for the thing 🤗

[Pollen Robotics @pollenrobotics]

We built a small biped robot you can teach new tricks to.

Train it in simulation, run it on the real thing. Meet Microduck 🦆

$399, shipping before Christmas.

https://t.co/RflJlIUwOu
https://t.co/kcoCKdAKfu

[Soumith Chintala @soumithchintala]

once you know the tasks you care about,
customization >> general

[Vishnu Samavedula @svrvishnu]

🫣😳

[Forbes @Forbes]

Nvidia Has Reportedly Agreed To Buy AI Model Hosting Platform Hugging Face For $13 Billion
https://t.co/42v8V53sSY

[Ramin @ramin_m_h]

coming to Pipette!
https://t.co/dUdBzo41nh

[kieran @KieranPDennis]

Really important observation from Chamath in Friday’s All-In.

Open source models inside frontier harnesses are INCREDIBLE.

Setting that up is also about to get a ton easier 😉

[Vishnu Samavedula @svrvishnu]

Looking to implement intelligence on edge? Check out @ArtificialAnlys 's new testing results.

[Artificial Analysis @ArtificialAnlys]

Which model should you run on the iPhone 17 Pro?

Announcing our new intelligence and inference testing for small models on mobile devices: independent measurement of how capable small models are in typical on-device tasks, and how they perform on popular phones - in partnership with @liquidai

We have partnered with @liquidai to deliver mobile device inference benchmarking, covering a range of models in 4-bit or lower precision on the iPhone 17 Pro and Galaxy S26 Ultra

We’re publishing our phone-scale intelligence evaluation results in combination with @liquidai's inference performance benchmarks to give users and developers a holistic view of how small models are performing on phones

Inference benchmarking is conducted in a controlled environment using Liquid AI’s inference benchmarking software, which Artificial Analysis has examined and is open-sourced on Liquid AI’s GitHub. Results cover end-to-end generation time, output speed, peak memory usage and other metrics. The inference benchmarking app, ‘Pipette’, is available to download for free on iOS and Android, allowing users to test a variety of models on their own devices

We are ranking phone-scale model intelligence based on each model’s average score in five evaluations chosen for the task-based work these models do in practice: BFCL, IFBench, AA-Omniscience, GPQA Diamond and MATH-500. These evaluations are run by Artificial Analysis using our independent methodology. By default, we limit models to 16K context on each of these evaluations, representing the lack of memory space for significant KV cache on mobile devices. This leads to some intelligent but verbose models dipping in relative score - they were not designed for the constraints that phone memory imposes on token use

We are defining our portable device category as including models that fit within 8 GB of memory after quantization, including KV cache, at 8K context

We expect both the intelligence and inference benchmarks to evolve over time, as new models, devices, inference frameworks and quantization techniques are released. Our pages will remain up to date with these new additions

Initial results:

➤ Nanbeige4.2-3B and LFM2.5-2.6B share the top average evaluation score at 63 (with a 16K context limit), ahead of Ornith-1.0-9B at 62 and Qwen3.5 9B (Reasoning) at 61. LFM2.5-2.6B achieves its score more efficiently: on an iPhone 17 Pro it answers a standard 1,024-token prompt in 8.0s using 2.3 GB of memory, against 21.4s and 4.0 GB for Nanbeige4.2-3B and 25+ seconds and 6.9 GB for the two 9B models

➤ The 16K context limit shapes the leaderboard: as an example, Qwen3.5 9B (Reasoning) spends 74.5M output tokens across one pass of the benchmark set, hitting the 16K limit on 29% of its generations and landing in fourth place overall. With the limit raised to 64K, Ling 3.0 Tiny takes first place with a score of 66, ahead of Nanbeige4.2-3B (65), Qwen3.5 9B (Reasoning, 64), and LFM2.5 2.6B (64). But a 64K window does not fit in mobile phone memory, and at 55 output tokens/s on an iPhone, generating 64K tokens could mean a 20+ minute wait and a lot of battery use. This is why our primary results are capped at 16K, but we're also publishing a set of results capped at 64K, and another capped at one minute of generation time

➤ The speed-intelligence Pareto frontier is short: six models are unbeaten on both intelligence and speed on an iPhone 17 Pro: LFM2.5-230M (27 at 0.9s), MiniCPM5-1B (45 at 2.9s), LFM2.5-8B-A1B (58 at 5.7s), Ling 3.0 Tiny (59 at 5.7s), LFM2.5-2.6B (63 at 8.0s) and Nanbeige4.2-3B (63 at 21.4s). LFM2.5-8B-A1B and Ling 3.0 Tiny are mixture-of-experts models that activate ~1B parameters per token, which is how they answer in under 6s with 8B-class weights

➤ Leading models have opposite strengths: Nanbeige4.2-3B is the most balanced (76% on BFCL, 96% on MATH-500, 67% on GPQA Diamond); Qwen3.5 9B (Non-reasoning) is the strongest tool caller (77% on BFCL) and scientific reasoner (79% on GPQA Diamond); LFM2.5-2.6B follows instructions best of any model measured on the iPhone (59% on IFBench), clears 90% on MATH-500 and hallucinates far less on AA-Omniscience (79% non-hallucination, against 33% for Nanbeige4.2-3B and 24% for Qwen3.5 9B (Reasoning))

More details below in thread ⬇️

[Vishnu Samavedula @svrvishnu]

Introducing Pipette ! An open source benchmarking platform for on-device intelligence.

[Liquid AI @liquidai]

Today we release Pipette, a model evaluation suite for on-device intelligence, in partnership with @ArtificialAnlys

Most benchmarking platforms are optimized to measure core capabilities and speed profile of foundation models served in the cloud. Pipette gives the field a common, reproducible way to measure the quality, speed, latency, and memory use of AI models on devices such as phones, laptops, PCs, AI boxes, and embedded hardware.

> Pipette is open source
> In Pipette, models get compated as model + quantization + runtime + device from one interface.
> It comes with a warehouse of verified benchmark results, currently with 10k+ results across 35 model classes, 7 quants, llama.cpp runtimes, and 4 devices.
> Pipette is a dynamic platform, allowing new contributions from day one, adding new devices, runtimes, model families, and quantization levels.

🧵

[Vishnu Samavedula @svrvishnu]

Try it now !!

[Maxime Labonne @maximelabonne]

LFM2.5-2.6B and LFM2.5-Embedding-350M are now available for free on @OpenRouter 🚀

[Liquid AI @liquidai]

Today, we release updated 4-bit checkpoints for LFM2.5-230M, LFM2.5-350M, LFM2.5-1.2B-Instruct, and LFM2.5-2.6B trained with Quantization-Aware Distillation (QAD). These checkpoints recover accuracy lost to 4-bit quantization while retaining the low memory footprint and high decode throughput of the Q4_0 format.

All four checkpoints reach roughly 97% of their BF16 averages.

🧵

[OpenMed @OpenMed_AI]

We gave LFM2.5-VL-3B from @LiquidAI a generated skin-like image and let it choose tools.

It mapped 6 regions, drew the contours, measured L04 at 8.8 × 8.1 mm, then picked what to review first.

Local on a Mac Studio. Visual review, not diagnosis. 🧵

[Ramin @ramin_m_h]

@jolfss @liquidai Let's look at throughput stoo. LFMs even with 5x-7x more active params to the openai moel are 3-6x faster. i just ran these on my mac just now.