Laguna XS 2.1
Release in the Laguna family · version XS 2.1
Laguna XS 2.1 is a mixture-of-experts model with 33B total and 3B active parameters per token, built for agentic coding on a local machine. Poolside describes it as an upgraded version of Laguna XS.2 with the same architecture. It supports interleaved reasoning between tool calls and has a 262,144-token context window.15
- Model hub: Model card (external site: huggingface.co)
- License: License (OpenMDW-1.1) (external site: huggingface.co)
- Release notes: Release announcement (external site: poolside.ai)
- Paper: Laguna M.1/XS.2 technical report (external site: poolside.ai)
Availability and license
Overall availability
Weights download from Hugging Face without an access gate. Also served through the Poolside API and OpenRouter.15
Availability is separate from permission: read the license before using or redistributing.
OpenMDW-1.1 grants broad use rights and places no restrictions on outputs. It requires keeping the license and notices when redistributing, and ends the rights of anyone who sues claiming the materials infringe a patent or copyright. The predecessor, Laguna XS.2, is published under Apache 2.0. The model card asks users to follow Poolside's Acceptable Use Policy.214
Model-disclosure tier
The model parameters for this release can be downloaded by the public. License terms may still restrict use, redistribution, or commercial use.
The weights are under a license that is not on the rubric's OSI-approved list. Read its terms before use.
Public materials checklist
| Item | Status | Notes and evidence |
|---|---|---|
| WeightsCan the general public download the model parameters for this release? | Public | BF16 weights are in the ungated repository; FP8, NVFP4, and INT4 variants are published as a collection.15 |
| Inference codeIs code for running the model published? | Public | The repository includes modeling code (modeling_laguna.py), and the model card documents vLLM, SGLang, Transformers, TensorRT-LLM, llama.cpp, and Ollama.13 |
| Training codeIs the code used to train the model published? | Unknown | Not assessed. |
| Training-data informationPublic = the training data itself can be obtained. Partial = composition or sources are documented without full access. | Partial | The model card points to the Laguna M.1/XS.2 technical report, which describes the XS.2 pre-training mix (web, code, curated educational and academic text, and synthetic data). Any data changes for XS 2.1 are not described, and no public data release was found.17 |
| Training recipeAre the training configuration and procedure documented in enough detail to follow? | Partial | The technical report documents XS.2 pre-training, supervised fine-tuning, and agentic reinforcement learning. The changes made for XS 2.1 are summarized, not specified.71 |
| Evaluation materialsPublic = evaluation code or prompts that let others re-run the evaluations are published. Partial = results only. | Partial | The model card publishes benchmark results with the harness, sampling settings, and sandbox resources used.1 |
What it is useful for
Run and use notes
- The model card documents support in vLLM, SGLang, Transformers, TensorRT-LLM, and llama.cpp, plus Ollama (with MLX) and mlx-lm for local use on a Mac. It notes a pending llama.cpp fix for empty output with Ollama on Apple Silicon.1
Organization context
Provenance and derivatives
The model card describes Laguna XS 2.1 as an upgraded version of Laguna XS.2. The technical report states that XS.2 was trained from scratch by Poolside.17
- Derived from: Laguna XS.2 (external site: huggingface.co) — Same organization; earlier release in the family.
Other releases in the Laguna family
- Laguna S 2.1Model-disclosure tier (USASI rubric v0.1): Open-weight
U.S. eligibility
Eligible · basis: U.S. headquarters
Developed and published by Poolside, Inc., which states that it is headquartered in San Francisco. See the poolside organization record for the assessment of its French presence. It updates Laguna XS.2, which Poolside's technical report says it trained from scratch.817
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.