Introducing Laguna S 2.1 — Poolside

Laguna S 2.1 Overview

Today we’re releasing Laguna S 2.1, a significant step forward in our development of models that pursue longer horizon work and make effective use of reasoning.

Laguna S 2.1 is a 118B total parameter Mixture-of-Experts (MoE) model with 8B activated parameters per token and supports a context window of up to 1M tokens in thinking and no-thinking modes. It went from the start of training to launch in under nine weeks, and on long-horizon coding benchmarks it holds its own against models many times its size. For every benchmark score we publish today, we are releasing full trajectories for every trial in the final evaluation set at trajectories.poolside.ai.

• Laguna S 2.1118B-A8B
• Tencent Hy3295B-A21B
• Inkling975B-A41B
• Nemotron 3 Ultra550B-A55B
• DeepSeek-V4-Pro-Max1.6T-A49B
• Kimi K32.8T-A50B
• Qwen 3.7 Max—
• Muse Spark 1.1—
• Claude Fable 5—

Terminal-Bench 2.1

Terminal-Bench 2.1 resolved tasks on Terminal-Bench 2.1.

A closer look at DeepSWE

The benchmarks above are all meaningful, and we're glad to be close to the frontier on them. On DeepSWE v1.1, Laguna S 2.1 scores 40.4 in thinking mode in pool harness.

Evaluation methodology

Evaluation of agent models is notoriously difficult due to the prevalence of reward hacking. We have previously written about reward hacking in leading benchmarks and our evaluations system and rigor.

Seeing the model work

Benchmark scores give a quantitative view into the model behavior, but to get a better intuitive understanding of how the model works, it’s useful to look into runs on real world tasks. We share three such tasks with unedited trajectories and commentary.

A browser engine from a blank folder

One of our favorite things about Laguna S 2.1 is its resourcefulness: It will find clever ways to get to the goal even if the direct path is not available. We asked it to build a browser engine from scratch; knowing it would be a challenge for Laguna to verify its work given its lack of vision capabilities.

// the verbatim prompt · reproduce it yourself
your job is it to build a simple browser engine (just html/css) in
javascript to demonstrate the capabilities of poolsides new "Laguna S"
model. the goal is to take render html snippets in a canvas like a real
browser. to demonstrate it the engine, build a self-contained single
page app that showcases a gallery of multiple html snippets and renders
them side by side.

Case study 2

Laguna S 2.1 is capable of meaningful engineering and research work. Over multiple hours of work, Laguna S 2.1 found and implemented multiple different optimizations in our agent harness, resulting in an overall speedup of 5.2%, and reducing memory allocation by ~70%.

Thinking effort

Laguna S 2.1 has two thinking modes: off and max (enabled by default). Max thinking lifts S 2.1’s score on Terminal-Bench 2.1 from 60.4% to 70.2% and on DeepSWE from 16.5% to 40.4%.

Limitations

We are excited about the capabilities of Laguna S 2.1 but there are known limitations we are working on for the next iteration:

• Harness overfitting
• Nested tool calls
• Longer than expected thinking duration/overthinking

What actually changed in this model

The changes made with Laguna S 2.1 weren’t just about scaling up the model size but improving the verification and persistence of solving tasks effectively.

Get started

Laguna S 2.1 is available on Hugging Face. You can run Laguna S 2.1 locally or through a variety of platforms like Baseten’s Model Library and OpenRouter.

If you want the base model for research or your own post-training, email models@poolside.ai.