← All Models

laguna-s-2.1

Public
Model
Revisions

Poolside's 118B MoE model with 8B active parameters, built for long-horizon agentic coding

95.6K Downloads

2 stars

Capabilities

Reasoning

Minimum system memory

71GB

Tags

118B
laguna
Report

README

Laguna S 2.1

Laguna S 2.1 is a 118B-parameter Mixture-of-Experts model with approximately 8B parameters activated per token, designed for agentic coding and long-horizon work.

Highlights

  • Agentic coding: optimized for software engineering, terminal tasks, and long-running tool workflows.
  • Efficient MoE architecture: routes each token through 10 of 256 experts plus one shared expert.
  • Native reasoning: supports interleaved and preserved thinking across tool calls.
  • Long context: supports up to 262,144 tokens.

Custom Fields

Special features defined by the model author

Enable Thinking

: boolean

(default=true)

Controls whether the model will think before replying

Preserve Thinking

: boolean

(default=false)

Preserves reasoning content from prior assistant turns

Sources

The underlying model files this model uses

Based on

🤗lmstudio-community/Laguna-S-2.1-GGUF→

GGUF