Poolside Laguna XS.2

Poolside Laguna XS.2 loads in 23 GB at Q4_K_M and is comfortable from 48 GB of memory, Macs included from the MacBook Pro M5 Pro 48GB. Below: the full memory math, the cheapest card that runs it well, and the fastest machine we track.

PARAMETERS
33B (3B active)
FORMAT
Q4_K_M
MIN MEMORY
36 GB
BEST FOR
Agentic coding, long-horizon software engineering

Memory math

Weights (Q4_K_M)
23 GB
+ KV cache (16k)
~4.0 GB
Total at 16k
~27.0 GB
Comfortable from
48 GB

KV = fp16 estimate (q8_0 cache roughly halves it). "Comfortable" = weights + KV within the engine's tiered budget (~70-85% of memory).

Hardware snapshot

Cheapest GPU
~25 tok/s est. — ~$3,847 used (as of 2026-07-31)
Fastest
~121 tok/s est.
Runs on a Mac?
~40 tok/s est.

More Laguna models

Frequently asked questions

How much memory does Poolside Laguna XS.2 need?

23 GB for the Q4_K_M weights, plus ~4.0 GB of KV-cache at 16k context — about 27.0 GB total. Comfortable from 48 GB of VRAM or unified memory.

Does Poolside Laguna XS.2 run on a Mac?

Yes — from the MacBook Pro M5 Pro 48GB (~40 tok/s est.). Unified memory means the RAM budget is the only limit.

What is the cheapest GPU for Poolside Laguna XS.2?

The AMD Ryzen AI Max+ 395 (Strix Halo) is the cheapest tracked card that runs Poolside Laguna XS.2 comfortably — ~25 tok/s est. at ~$3,847 used (as of 2026-07-31).

Cite this page

ModelFit: Poolside Laguna XS.2 — specs, memory math and hardware verdicts.
https://modelfit.io/models/laguna-xs2/ (dataset updated 2026-09-03, CC BY 4.0).