---
title: Laguna S 2.1
link: https://zml.ai/posts/laguna-s-2-1/
source: zml-ai
published: 2026-07-21T18:00:00Z
updated: 2026-07-21T18:00:00Z
first_seen: 2026-09-08T10:17:10.166315Z
summary: 'In close collaboration with Poolside, we’re happy to announce the support of Laguna S 2.1, their new open-weight model for agentic coding and long-horizon work. It is a 118B total parameters Mixture-of-Experts model with 8B activated parameters per token and supports a context window of up to 1M tokens, available in thinking and no-thinking modes. And it’s available today in ZML/LLMD with DFlash support. Most importantly (and as always with us), it’s available on all the supported platforms: NVIDIA, AMD, Intel, Google TPU and Apple Metal.'
content: extracted
html: 2026-07-21-laguna-s-2-1.html
images:
- source: https://zml.ai/posts/laguna-s-2-1/laguna-s-2-1-chart.svg
  original:
    file: 2026-07-21-laguna-s-2-1.image-1a349976835c.png
    width: 699
    height: 1102
  variants:
  - file: 2026-07-21-laguna-s-2-1.image-476a87f8698b.webp
    width: 320
    height: 504
  - file: 2026-07-21-laguna-s-2-1.image-866996f2b71d.webp
    width: 640
    height: 1009
  - file: 2026-07-21-laguna-s-2-1.image-9e58d16c012d.webp
    width: 699
    height: 1102
  color: '#000000'
---

[ZML - Model to Metal](https://zml.ai/)

- [Blog](https://zml.ai/posts/)
- [LLMD](https://zml.ai/llmd/)
- [GitHub 3.5K](https://github.com/zml/zml)
- [Docs](https://github.com/zml/zml/tree/master/docs)
- [Discord](https://discord.com/invite/puhdWt8gzZ)
- [Contact](https://zml.ai/cdn-cgi/l/email-protection#e088858c8c8fa09a8d8cce8189)

[Back to posts](https://zml.ai/posts/)

Article

Poolside's open-weight model for agentic coding and long-horizon work.

July 21, 2026 inference / performance / NVIDIA / AMD / TPU / Intel / Apple / Metal / poolside / laguna

In close collaboration with [Poolside](https://poolside.ai/), we’re happy to announce the support of [Laguna S 2.1](https://poolside.ai/blog/introducing-laguna-s-2-1), their new open-weight model for agentic coding and long-horizon work.

It is a **118B total parameters** Mixture-of-Experts model with **8B activated parameters** per token and supports a context window of up to **1M tokens**, available in **thinking** and **no-thinking** modes.

And it’s **available today** in ZML/LLMD **with DFlash support**. Most importantly (and as always with us), it’s available on all the supported platforms: **NVIDIA**, **AMD**, **Intel**, **Google TPU** and **Apple Metal**.

## Run it today

**NVIDIA**

```bash
docker run -p 8000:8000 --shm-size=256GB --gpus=all -e HF_TOKEN -it zmlai/llmd:cuda \
    --model=hf://poolside/Laguna-S-2.1 \
    --dflash-model=hf://poolside/Laguna-S-2.1-DFlash
```

**AMD**

```bash
docker run -p 8000:8000 --shm-size=256GB --device=/dev/kfd --device=/dev/dri -e HF_TOKEN -it zmlai/llmd:rocm \
    --model=hf://poolside/Laguna-S-2.1 \
    --dflash-model=hf://poolside/Laguna-S-2.1-DFlash
```

**Intel**

```bash
docker run -p 8000:8000 --device=/dev/dri -e HF_TOKEN -it zmlai/llmd:oneapi \
    --model=hf://poolside/Laguna-S-2.1 \
    --dflash-model=hf://poolside/Laguna-S-2.1-DFlash
```

**Google TPU**

```bash
docker run --net=host --privileged -e HF_TOKEN -it zmlai/llmd:tpu \
    --model=hf://poolside/Laguna-S-2.1 \
    --dflash-model=hf://poolside/Laguna-S-2.1-DFlash
```

**Apple Metal**

```bash
brew install zml/zml/llmd
llmd --model=hf://poolside/Laguna-S-2.1 --dflash-model=hf://poolside/Laguna-S-2.1-DFlash
```

## Benchmarks

*It’s a good model, sir.*

![](https://zml.ai/posts/laguna-s-2-1/laguna-s-2-1-chart.svg)