Ollama · laguna-s-2.1

![laguna-s-2-1-banner.svg](/assets/library/laguna-s-2.1/0f2f47dd-ff7b-47ad-85dd-83c15d521055) # Laguna S 2.1 ## Overview Laguna S 2.1 is a Mixture-of-Experts model built for agentic coding and extend…

ollama/laguna-s-2.1

![laguna-s-2-1-banner.svg](/assets/library/laguna-s-2.1/0f2f47dd-ff7b-47ad-85dd-83c15d521055)

# Laguna S 2.1

## Overview

Laguna S 2.1 is a Mixture-of-Experts model built for agentic coding and extended reasoning. It is the larger sibling of Laguna XS 2.1, trading a bigger memory footprint for stronger coding and reasoning performance while keeping a small activated-parameter count so it stays practical to run on a single high-memory machine.

## Key Specifications

- **Total Parameters**: 118B (8B activated per token) - **Architecture**: 48 layers with mixed attention (12 global, 36 sliding window at 512) - **Experts**: 256 routed (top-10 per token) with 1 shared expert - **Context Window**: 256K tokens - **License**: OpenMDW-1.1

## Available Model Variants

| Tag | Size | Context | Type | |-----|------|---------|------| | laguna-s-2.1:latest | 75GB | 256K | Text | | laguna-s-2.1:q4_k_m | 75GB | 256K | Text | | laguna-s-2.1:q8_0 | 128GB | 256K | Text | | laguna-s-2.1:f16 | 235GB | 256K | Text |

## Benchmark Results

<!-- TODO: fill from official Laguna S 2.1 results before publishing. Do not reuse XS numbers. -->

## Technical Highlights

Sigmoid expert gating with normalised expert weights and per-layer rotary scales. Attention alternates one global layer for every three sliding-window layers, with grouped-query attention (8 KV heads) keeping the KV cache small at long context. Supports tool calling and interleaved thinking, with thinking enabled or disabled per request.

## Usage

```bash ollama run laguna-s-2.1 ```

Use our coding agent, [pool](https://github.com/poolsideai/pool), for the best experience.

```bash POOLSIDE_STANDALONE_CONTEXT_LENGTH=262144 ollama launch pool --model laguna-s-2.1 ```

Tool calling and thinking work through Ollama's built-in template:

```bash # thinking on a single request curl http://localhost:11434/api/chat -d '{ "model": "laguna-s-2.1", "messages": [{"role": "user", "content": "If 3x+6=21, what is x?"}], "think": true }' ```

### Hardware and settings

The q4_k_m weights need roughly 75GB of memory, so a 128GB unified-memory machine (Apple Silicon or DGX Spark) is the practical minimum, and measured single-stream decode is around 13-18 tokens/s on that class of hardware. It also runs on datacenter GPUs such as the NVIDIA H200, where far higher memory bandwidth gives correspondingly faster generation.

Two settings matter in practice:

- **Increase the load timeout.** A cold start reads 75GB from disk, which can exceed Ollama's default 5 minute model load timeout on slower storage. Set `OLLAMA_LOAD_TIMEOUT=20m` if the first run times out. - **Raise `num_ctx` deliberately.** Ollama defaults to a small context; long context costs memory on top of the 75GB of weights, so raise it only as far as your machine allows.

Thinking consumes output tokens before the final answer appears, so allow a generous `num_predict` on reasoning prompts or the reply can be cut off mid-thought.

## Notes

Verified on macOS (Apple Silicon, Metal), NVIDIA DGX Spark (GB10), and NVIDIA H200, with Ollama 0.31 and 0.32.

Specifications & capabilities

Published model ID
laguna-s-2.1

Deployments & artifacts

1–8 / 8

ollama

source id
ollama
execution mode
local
parameter count is approximate
false
quantization level
nvfp4
context window
1000000
context window is approximate
true
artifact size bytes
68419655521
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter size
117.6B
parameter count
117600000000
parameter count is approximate
true
quantization level
F16
context window
256000
context window is approximate
true
artifact size bytes
235202260871
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter size
117.6B
parameter count
117600000000
parameter count is approximate
true
quantization level
Q8_0
context window
256000
context window is approximate
true
artifact size bytes
96031832391
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter count is approximate
false
context window
1000000
context window is approximate
true
artifact size bytes
237361330061
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter count is approximate
false
quantization level
mxfp8
context window
1000000
context window is approximate
true
artifact size bytes
135840232474
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter count is approximate
false
quantization level
nvfp4
context window
1000000
context window is approximate
true
artifact size bytes
66189691080
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter size
117.6B
parameter count
117600000000
parameter count is approximate
true
quantization level
Q8_0
context window
256000
context window is approximate
true
artifact size bytes
96031832391
updated at
2026-11-16T08:10:06.079865+00:00

ollama

source id
ollama
execution mode
local
parameter size
117.6B
parameter count
117600000000
parameter count is approximate
true
quantization level
Q8_0
context window
256000
context window is approximate
true
artifact size bytes
127650803591
updated at
2026-11-16T08:10:06.079865+00:00

Original sources

1–2 / 2

Ollama · Model page ↗

Ollama · Catalog ↗

Source links

1–1 / 1

Explore more models

Models

Loading
EVIDENCE / COMPARISON

Benchmarks

Explore every benchmark

Loading

These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

Install TensorTrace