Loading video...

Video Failed to Load

Go Home

This is a 1.7B param model btw, using the browser, making html files, and managing itself in Pi. This model hits 4o levels on most benchmarks lol Great for workflow automations.

121,160 views • 5 days ago •via X (Twitter)

32 Comments

Jaro's profile picture
Jaro5 days ago

Yeah this lfm model is amazing - works on my android with 3gb ram at 20t/s

David Valerio Gilmore's profile picture
David Valerio Gilmore5 days ago

nice - I hadn't tried the new LFMs with more agentic capabilities. going to look into training hybrid edge/cloud semantic routing for it in our next ARC checkpoint

Dominik's profile picture
Dominik5 days ago

4o levels? So great at causing psychosis but not much else?

g023's profile picture
g0235 days ago

pretty epic. Also the 2B with vision punches way above its weight

Dragos Roua's profile picture
Dragos Roua5 days ago

You can even use it on a Raspberry Pi: It’s the stock one, without webGPU kernels.

Jeramy's profile picture
Jeramy5 days ago

I’ve been experimenting with this one too. For laptops without great chips, it’s pretty amazing what it can do!

srujan's profile picture
srujan5 days ago

heard a bunch of terrible stuff about lfm3, but glad to see someone doing some good stuff with this will def look into lfm3 automation into a new agent focused browser im working on

n0geegee's profile picture
n0geegee5 days ago

tried it, trashed it

Hyre's profile picture
Hyre5 days ago

a 1.7B model writing its own html files would’ve sounded made up last year

johnny smitthy's profile picture
johnny smitthy5 days ago

What's the model?

M M's profile picture
M M5 days ago

No local model matches 4o in anything.

The Singularity Layer's profile picture
The Singularity Layer5 days ago

Let's add this on Singularity Network.

Andrew Rose's profile picture
Andrew Rose5 days ago

This is the size I care about for office automations. My 14B on a 3080 runs around 18 tok/s with some layers on CPU, a 1.7B that can drive a browser leaves the whole card for context. How's it hold up on longer multi-step stuff?

𝖄𝖊𝖓's profile picture
𝖄𝖊𝖓5 days ago

Yup, my poor man’s local agent 🥹 works like a charm👌

A Concerned Human's profile picture
A Concerned Human5 days ago

It's pretty good but it's quality of work could use work

Alade Yessoufou's profile picture
Alade Yessoufou5 days ago

Which model is that ?

Juan Carlos Vez Vázquez 👣 Artesano en Mk y Ventas's profile picture
Juan Carlos Vez Vázquez 👣 Artesano en Mk y Ventas5 days ago

Awesome!

Akash Sardar's profile picture
Akash Sardar5 days ago

Lfm 3 my guess i have tried it too

Donovan Baney's profile picture
Donovan Baney5 days ago

It is certainly one of the greatest models.

Adi R 🇺🇲🇮🇱🇺🇦💙🌎👽's profile picture
Adi R 🇺🇲🇮🇱🇺🇦💙🌎👽5 days ago

It's Mac browser I assume? My iPhone browser refuses to run even 325m param model - site. Runs great on Mac though. Feedback is most welcome!

Velsa 🌹💙's profile picture
Velsa 🌹💙5 days ago

fiquei curioso, essa interface bem feita, é de qual app? é nativa?

Nick's profile picture
Nick5 days ago

Gonna go check it out. AI for exploration is awesome.

ALUCAB 🇪🇺's profile picture
ALUCAB 🇪🇺5 days ago

this is great. any writeup or link i can read ? I am trying to run it in collaboration with picoclaw for example but it loses context also for simple action requests and conversations

Taylor W's profile picture
Taylor W5 days ago

Demo videos need bass-heavy instrumental dance soundtracks for me to take them seriously.

Jonny D's profile picture
Jonny D5 days ago

how do you recommend using a model like this?

Aercrist's profile picture
Aercrist4 days ago

How do you build custom silicon for Liquid AI’s LFM3-1.7B? The tri-hybrid architecture (Conv + Sliding Window + Global Attention) requires a completely different ASIC design than pure transformers. Here’s how to architect an edge AI chip for it.

Fin Boots's profile picture
Fin Boots5 days ago

Pretty insane this is even possible. How’s tool calling capability?

Ash's profile picture
Ash5 days ago

Liquid cooked with this one

Sven Meyer's profile picture
Sven Meyer5 days ago

on what inference engine do you run it ?

E1ev3n.'s profile picture
E1ev3n.5 days ago

The browser part is what got me. I run a small model on my own machine for the boring jobs, and the moment it stopped needing the cloud my whole workflow got calmer. A 1.7B model that makes its own HTML files and manages itself is the first time the small end of the scale has felt like the useful end to me.

落落's profile picture
落落5 days ago

Can it reliably run locally on a Pi?

Piotr Pabis's profile picture
Piotr Pabis5 days ago

And it’s amazing at vision, I was completely stunned

Related Videos