Skip to content
localmodel.run

Text model · Laguna

LA Laguna S 2.1: RAM and VRAM requirements

Laguna S 2.1 needs about 92.8 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~89.44 GB to download; KV cache and overhead add the rest), or about 123.3 GB at Q8_0. The lightest hardware that runs it is Apple M4 Max (128GB).

Laguna family · 118B params (Mixture-of-Experts: activates only 8B of 118B params per token, so generation is faster than the total size suggests) · released Jul 2026.

Shopping for hardware? See what runs Laguna S 2.1 →

License OpenMDW-1.1 · Commercial OK ↓ 81.6K/mo ♥ 897 on HuggingFace
Q4_K_M GGUF
89.44 GB
Q8_0 GGUF
119.91 GB
Memory @ Q4 (4k)
~92.8 GB
Context
1000 k

Q4_K_M and Q8_0 are GGUF download (file) sizes. Memory @ Q4 includes the KV cache and overhead, so it runs larger than the file on disk.

Will it run on your device?

Laguna S 2.1 runs on 4 of 40 tracked devices at Q4_K_M.

1 run well 3 tight fit 36 too small
Fit check Q4_K_M
No, not enough memory
needs 92.8 GB usable 10.5 GB
See the full breakdown

Memory breakdown

Weights (Q4_K_M)89.44 GB
+
KV cache (4k)2.6 GB
+
Overhead0.8 GB
=
Total92.8 GB

How context length changes it

4k context ~92.8 GB 32k context ~111.1 GB 128k context ~173.6 GB

Longer context grows the KV cache, which for Laguna S 2.1 is sized by its 8B active params, not the full 118B. It needs ~92.8 GB at 4k and ~173.6 GB at 128k.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 49.4 GB est
Q3_K_M 57.7 GB est
Q4_K_M (default) 89.44 GB
Q5_K_M 84.1 GB est
Q6_K 96.8 GB est
Q8_0 119.91 GB
FP16 219.05 GB

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama
$ ollama run laguna-s-2.1:latest
llama.cpp
$ llama-cli -hf poolside/Laguna-S-2.1-GGUF:Q4_K_M
LM Studio
$ lms get poolside/Laguna-S-2.1-GGUF

Which devices can run Laguna S 2.1?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

NVIDIA GPUs

AMD GPUs

Same job, different size

Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.

Next size down gpt-oss 120B 117B Next size up Nemotron 3 Super 120B-A12B 120B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does Laguna S 2.1 need?

At Q4_K_M, Laguna S 2.1 needs about 92.8 GB (weights ~89.44 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~123.3 GB.

What is the Q4_K_M GGUF file size of Laguna S 2.1?

The Q4_K_M GGUF file is about 89.44 GB to download, and the Q8_0 GGUF is about 119.91 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~92.8 GB of memory at Q4_K_M.

Can Laguna S 2.1 run on a laptop?

Laguna S 2.1 is large; you need a high-memory Mac or multi-GPU setup at Q4_K_M.

Is Laguna S 2.1 cheaper to run because it is a MoE model?

It is faster, not lighter. Laguna S 2.1 activates only 8B of 118B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 118B.

Can I use Laguna S 2.1 commercially?

Yes. Laguna S 2.1 is licensed OpenMDW-1.1, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Laguna S 2.1's memory and quant figures.

Poolside AI's coding-agent flagship: 118B total / 8B active MoE, built for long-horizon agentic work (70.2% on Terminal-Bench 2.1). 1M native context. Sizes from Poolside's own official GGUF repo, cross-confirmed against the Ollama tag listing (exact match: 96/128/235GB). Linux Foundation's OpenMDW-1.1 license: permissive, free commercial use.

Sources

Last validated 2026-08-03. Memory figures are estimates. See methodology.