Laguna S 2.1

https://poolside.ai/blog/introducing-laguna-s-2-1

by rexledesma • 1 day ago

395 78 1 day ago

poolside.ai

tosh• 1 day ago

open weights and

similar performance to deepseek v4, inkling at size of nemotron 3 super (!)

iraldir• 1 day ago

Amazing model at this size if true, that's quite crazy!

megavon3• 1 day ago

This is INSANE. How did they do this?

eisokant• 1 day ago

"What we've done in this model is not necessarily add more intelligence, but improve the behaviors that lead to a more capable model: more verification, less taking things for granted, not declaring victory early, and being more persistent.”

https://poolside.ai/assets/laguna/laguna-m1-xs2-technical-re...

Lwerewolf7• 1 day ago

Almost like a built-in heavyweight harness.

kamranjon3• 1 day ago

"It went from the start of training to launch in under nine weeks..."

This is pretty impressive.

aitchnyu• 14 hours ago

Are all AI labs in Google's weight class in crawling and ranking the Web's content? I know OpenAI has contractor subject matter experts in all topics.

Iolaum• 1 day ago

Model Looks amazing!

Even more important, subjectively, is that this model will run very well on Strix Halo (e.g. Framework Desktop), DGX Spark kinds of devices. Looking forward to Unsloth dynamic mtp quants.

P.S. Looking at the HF release they already offer Q4_K_M and DFlash drafter for speculative decoding!

verdverm4• 1 day ago

I hope all models going forward come with a dflash drafter so we don't have to train one up separately.

SwellJoe• 1 day ago

This is exactly the kind of model that's been needed in the middle. Realistically self-hosted, Good Enough intelligence, MoE so it's fast on limited bandwidth systems like Strix Halo and DGX Spark.

For a while there's been nothing to run on my Strix Halo that's notably better than what I can run on my dual 32GB GPU desktop (Gemma 4 or Qwen 3.6 dense models), but this seems likely to be the step up in size that actually works better than those.

mandelken• 59 minutes ago

On Strix Halo, have a look at the work antirez is doing with dwarfstar (antirez/ds4) for deepseek v4 flash at reasonable speeds and quality.

river_otter4• 1 day ago

I love this. Is it possible to give a feel of how this stacks up to the good old Opus 4.5 in coding quality? For me that was the turning point where agentic coding in Claude Code etc became usable. Have we hit that threshold?

megavon3• 1 day ago

Having played with it for like 3 hours now....I'm probably moving from CC to this

river_otter4• 1 day ago

I am about 1 hour into using it with pi.dev. Do you have thinking on high? It is doing good but at one point i had to stop it and say 'you're overthinking this' haha

fingerprinter• 1 day ago

One hour in, no more Codex for me. This thing rips.

kamranjon3• 1 day ago

Whoa whoa whoa, 118b params, 8b active MOE, long context reasoning, open weights - music to my ears. Hadn't heard of this lab before but I am very excited, will definitely try this out tomorrow - this is a real sweet spot I think in terms of model size and performance.

svclaws• 1 day ago

If the numbers are legitimate then our prayers have been heard

Lwerewolf7• 1 day ago

Testing it now. At the very least, competitive with DS4-Flash indeed. On my small (and per Sol's words, _very_ semantically dense) C test codebase, it found things that only gpt-5.2 managed to find back in the day, but also made a stupidly incorrect initial observation that a memfd_create()/mmap was used for IPC (funnily enough - sol missed that as well in its review, until I pointed it out). Re: the claims vs deepseek v4 - both flash and pro are expected to get a "general availability" release very soon (i.e. well-"post-trained"), so things can change in a... well, flash, as per usual in the current environment.

Anyways, keep 'em coming.

ilc• 1 day ago

What harness/quant did you use for testing?

Lwerewolf7• 1 day ago

nvfp4 mlx, literally barebones pi.

edit: on bigger tests, got it to loop pretty easily unfortunately, probably local settings.

mchusma• 1 day ago

Incredible. This is definitely the launch of the day. Just crushing Google's releases.

The pricing here is incredible. This is the first US release that's competitive with DeepSeek V4 Flash. Very excited about this.

river_otter4• 1 day ago

Hey, this model is not a joke! Exciting, we already got a usable PR of work out of it.

https://github.com/mozilla-ai/otari/pull/348

mft_3• 1 day ago

Looks impressive, and this size fits achievable home hardware.

That said, if someone would kindly quantise this down for the 64GB paupers, that would be appreciated. (I know there’s likely degradation, but some people reported good results with a 2 bit version of Qwen 3.5 122B, and this is starting from a higher point. Would be interesting to try, at least.)

Edit: someone in the process of doing so: https://huggingface.co/vcruz305/Laguna-S-2.1-GGUF

verdverm4• 1 day ago

The tool I've been using, llm-compressor, can quant models that do not fit in memory (use the sequential pipeline)

https://github.com/vllm-project/llm-compressor

my setup to help you on your way: https://github.com/verdverm/quantr

Though it seems these will not be needed as Poolside has published quants & dflash with their models.

alfiedotwtf3• 15 hours ago

Nice! Do you know of any tools that do this for tensorrt models?

yogeshp• 1 day ago

They have also published smaller 33B model called Laguna XS 2.1, its Q4 gguf is 20GB.

https://huggingface.co/poolside/Laguna-XS-2.1-GGUF/tree/main

mft_3• 1 day ago

Thanks for flagging. From the few benchmarks I can find, it looks there or thereabouts with Qwen 3.6-35B-A3B, or maybe a touch below. I'm interested to compare a model that is a big jump larger with pretty impressive benchmarks, but more heavily quantized to fit.

Archit3ch• 23 hours ago

if someone would kindly quantise this down for the 64GB paupers

The Q4_K_M is 75GB. I'm exactly at 64GB, and I wouldn't quantize it further. Instead, do partial weight residency and stream the rest from SSD.

nunodonato• 15 hours ago

unsloth is always here to save the day https://huggingface.co/unsloth/Laguna-S-2.1-GGUF

andy993• 14 hours ago

Seems it’s not fully supported in mainline llama.cpp yet https://github.com/ggml-org/llama.cpp/pull/25165

In the huggingface link they mention building for CPU and for CUDA, does anyone know if that means it wouldn’t be possible to build targeting Vulkan?

mft_3• 9 hours ago

This is interesting: https://github.com/tanishq-dubey/macos-laguna-s2.1

Someone has benchmarked a wide range of different quants.

carimura• 1 day ago

Congrats Poolside team!!

docheinestages• 1 day ago

Any estimates of the performance (prompt processing and decoding tokens/s) on consumer hardware like Macbook Pro M-series?

benjiro29• 1 day ago

!! Be careful when testing the model.

A lot of people are testing it, and reporting disappointed results / benchmaxxxing claim. But do not realize that thinking has a issue with the default configuration.

Important - make sure that THINKING is enabled. By default it wasn't although I was passing the flag --default-chat-template-kwargs '{"enable_thinking": true}' in vllm recipe. The generation_config.json file that is included has by default max_new_tokens as 32k which seems to be cutting off thinking altogether so increase it. At first I was very disappointed with the output I was seeing, but once thinking is enabled, the code quality seems to be MUCH better. More real world testing to be done.

https://www.reddit.com/r/LocalLLaMA/comments/1v2pg99/laguna_...

voxgen• 1 day ago

Even the official provider on OpenRouter seems to have this issue. Hope it's an easy fix for them.

nshotton• 1 day ago

Thanks for posting this, it made a huge difference tweaking the recipe.

d2p• 15 hours ago

Looks like the default chat template was updated on HF to enable this by default shortly after you posted this :-)

danr4• 1 day ago

holy shit its accelerating fast

loolhahalmao• 1 day ago

happy the US has some counterweights to the Chinese labs, just need about half a dozen more.

aubanel• 1 day ago

Really impressive signal that this 128B model can beat DeepSeek V4 (1.6T) on most coding benchmarks!

Also, I really like Poolside's habit to compare not only to other top models in its weight class (others don't do it, looking at you Mistral), but also to the very top open-weight models, even much bigger ones like the 2.5T Kimi-K3!

spelk• 1 day ago

This is fantastic work, really impressive is an understatement. I really hope this sets a new DeepSeek-esque standard and starts another the death knell for companies continuing to cosplay as frontier labs (like Cohere).

_mrinalwadhwa_2• 1 day ago

Has anyone tried it on a mac yet?

_mrinalwadhwa_2• 19 hours ago

Was able to run it on Apple M3 Max (128 GB)

host: Apple M3 Max, 128 GB model: Laguna-S-2.1, 118B-A8B MoE, Q4_K_M (75 GB), DFlash speculative decoding server: http://127.0.0.1:8000, llama.cpp, ctx 64K, 8-bit KV cache

  mode: max thinking
   #  tokens   tok/s  dflash
   1     600    14.4     11%
   2     600    26.1     27%
   3     600    17.8     18%
   4     600    14.0     16%
   5     600     9.3     15%
  --------------------------------
  median  14.4   mean  16.3   min   9.3   max  26.1   tok/s

  mode: no thinking
   #  tokens   tok/s  dflash
   1     190    10.0     20%
   2     109    26.7     65%
   3      95    29.6     72%
   4      93    32.8     81%
   5     382    14.0     30%
  --------------------------------
  median  26.7   mean  22.6   min  10.0   max  32.8   tok/s

axus• 1 day ago

You will not, directly or indirectly: copy, modify, or create derivative works of the Site or any Product;

verdverm4• 21 hours ago

initial impressions, great model for coding, probably swapping it out for qwen 27b for a while to long-term test, more sycophantic than any I've run locally myself

drob518• 20 hours ago

Immediate reaction is that it seems to be a bit behind Meta Muse Spark 1.1 performance at approximately the Deepseek v4 Flash price point. That's quite good given Muse Spark benchmarks a lot better than Deepseek v4 Flash (assuming benchmarks mean anything, which they don't).

polski-g• 20 hours ago

This thing is great, twice as fast as DS4Flash and slightly smarter too. I swapped most of my sub-agents to this model.

platinumrad• 20 hours ago

Does it refuse to work on "cyber"?

kouteiheika• 12 hours ago

It's an open-weight model so it literally doesn't matter whether it refuses by default or not, because it's pretty trivial to uncensor[1] any open-weight model and make it not refuse.

[1] https://github.com/p-e-w/heretic

markasoftware• 19 hours ago

have only tested a few prompts, but it failed my favorite non-coding question that dsv4 flash aces. Benchmarks look excellent though (don't they always!)

luciana1u• 19 hours ago

64GB RAM used to mean you could run three VMs and a database. now it means 'maybe the 120B model fits if I close everything else'

resonious• 18 hours ago

How much does it cost? I even made an account and I cannot find pricing anywhere...

asar• 15 hours ago

In / Out Price

$0,10 / $0,20per 1M

from openrouter

alfiedotwtf3• 15 hours ago

Anyone know if this is supported in ik_llama.cpp?

suprjami• 14 hours ago

Just merged into main llama.cpp a few hours ago:

https://github.com/ggml-org/llama.cpp/pull/25165

throwdbaaway• 14 hours ago

It works, thanks to https://github.com/ikawrakow/ik_llama.cpp/pull/1911, which got merged in early June. However, there might still be some issue with the chat template.

alfiedotwtf3• 8 hours ago

Awesome, thanks!

Edit: it amazes me how fast ik_llama.cpp moves

reindeer2• 12 hours ago

I've been following work on the second-order effects that ripple through the system for a while. This is the first treatment I've seen that the framing reveals an assumption that isn't explicitly defended.

khurs• 11 hours ago

chat as listed on that page: https://chat.poolside.ai

mark_l_watson• 11 hours ago

For me, poolside.ai “came out of nowhere” a week or so ago when I discovered their local coding harness ‘pool’ and their smaller 33G MOE model that runs fast and is effective on my old 32G mac mini. Really good work!!

I need to evaluate their large hosted model.

samelldev• 9 hours ago

impressive benchmarks for the active params. i'll load it up and test

Laguna S 2.1

Related Stories