other
33B parameters
Commercial OK
Reviewed July 2026

Laguna XS 2.1

Laguna XS 2.1, released July 2, 2026, is Poolside's updated agentic-coding model (33B total, ~3B active MoE, 40 layers, 256 experts). It switched licensing from Apache-2.0 (the original Laguna XS.2) to OpenMDW-1.1, a permissive license co-developed with NVIDIA and the Linux Foundation specifically for AI model weights.

License: OpenMDW-1.1·Released Jul 2, 2026·Context: 262,144 tokens

Overview

Laguna XS 2.1, released July 2, 2026, is Poolside's updated agentic-coding model (33B total, ~3B active MoE, 40 layers, 256 experts). It switched licensing from Apache-2.0 (the original Laguna XS.2) to OpenMDW-1.1, a permissive license co-developed with NVIDIA and the Linux Foundation specifically for AI model weights.

Strengths

  • OpenMDW-1.1 license gives cleaner enterprise legal standing than a custom license
  • MoE (~3B active of 33B total) targets single-GPU agentic coding
  • Predecessor scored 44.5% SWE-Bench Pro, 30.1% Terminal-Bench 2.0 (vendor-reported)

Weaknesses

  • OpenMDW-1.1 is a new license (mid-2026) with less legal precedent than Apache-2.0/MIT
  • Predecessor (Laguna XS.2) was retired from Poolside's own API within 3 months of launch — fast iteration cadence, uncertain longevity per version
BLK · QUALITY BENCHMARKreviewed · raw logs

Reviewed quality benchmarks

First-party rows were run by RunLocalAI; reviewed community rows are labeled in the data. Every row links to the raw test-run log.

BenchmarkQuantRuntime / HardwareScoreRaw log
HumanEval+
tested 2026-07-21
Q4_K_M
ollama-0.32.1-vast5090
rtx-5090
87.2/100
Gist →

Q4_K_M note:First-party HumanEval+ on a rented Vast.ai RTX 5090 32GB instance via Ollama 0.32.1. Generation via evalplus_openai_generate.py, scoring via evalplus.evaluate (native Linux, no Windows shim needed). No DB credentials were placed on the rented host — results were pulled back and ingested from the operator's machine.

Want to verify? Every row links to its Gist with full stdout and stderr of the run. The runner script is in the public repo (scripts/run-humaneval-plus.ts) — reproducible end-to-end. Browse all coding scores at /benchmarks/coding.

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
Q4_K_M20.0 GB24 GB

Get the model

Ollama

One-line install

ollama run laguna-xs-2.1Read our Ollama review →

Hardware that runs this

Cards with enough VRAM for at least one quantization of Laguna XS 2.1.

Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Frequently asked

What's the minimum VRAM to run Laguna XS 2.1?

24GB of VRAM is enough to run Laguna XS 2.1 at the Q4_K_M quantization (file size 20.0 GB). Higher-quality quantizations need more.

Can I use Laguna XS 2.1 commercially?

Yes — Laguna XS 2.1 ships under the OpenMDW-1.1, which permits commercial use. Always read the license text before deployment.

What's the context length of Laguna XS 2.1?

Laguna XS 2.1 supports a context window of 262,144 tokens (about 262K).

How do I install Laguna XS 2.1 with Ollama?

Run `ollama pull laguna-xs-2.1` to download, then `ollama run laguna-xs-2.1` to start a chat session. The default quantization is Q4_K_M.

Source: Vendor official documentation

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.

Related — keep moving

Before you buy

Verify Laguna XS 2.1 runs on your specific hardware before committing money.