Language

Laguna S 2.1 (MoE)

Poolside Laguna S 2.1 — 118B-total agentic-coding MoE, 8B active, with a token-choice router using softplus gating over 256 routed experts plus 1 shared. Grouped-query attention with interleaved full and sliding-window layers, 1M context. Native interleaved thinking between tool calls; preserve reasoning blocks across turns. Requires llama.cpp b10087 or newer.

OpenMDW-1.1Source verified

Specification

Model ID

laguna-s-2.1

Family

laguna

Modality

Language

Parameters

118B (MoE, 8B active)

Context

1,048,576 tokens

Quantization

UD-Q4_K_XL

Weights

68.4 GB

Minimum RAM

80 GB

Source

Weights and provenance.

Repository

unsloth/Laguna-S-2.1-GGUF

Registry license

OpenMDW-1.1

Source license

openmdw-1.1

Access

Open — weights fetch without accepting additional terms.

Serve it

# Pull the weights onto a node
tenzro model download laguna-s-2.1

# Serve it
tenzro model serve laguna-s-2.1

# Or query the registry entry over JSON-RPC
curl https://rpc.tenzro.xyz \
  -X POST -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,
       "method":"tenzro_getModel",
       "params":[{"model_id":"laguna-s-2.1"}]}'

← All models