Skip to main content

Laguna S 2.1

PoolsideOpen WeightOpenMDW License v1.1 · Commercial OK

Description

Laguna S 2.1 is Poolside's most capable Laguna 2.1 model: an open-weight Mixture-of-Experts coding model with 118B total parameters and 8B activated per token (256 routed experts + 1 shared; 48 layers). It targets long-horizon and agentic coding with interleaved thinking, a 1,048,576-token context window in thinking and no-thinking modes, and text-to-text only modalities. Available as open weights on Hugging Face and via Poolside's API, OpenRouter, and other hosts (poolside/laguna-s-2.1).

Release Date
2026-07-21
Parameters
118.0B
Context Length
1.0M
Modalities
text

Capability Radar

60
general
60
coding
60
reasoning
51
scienceest.
60
agents
0
multimodal

Science is estimated from LLM Stats science scores or reasoning when no dedicated science benchmarks are available.

Rankings

No ranking data available

Benchmark Scores (LLM Stats)

(LLM Stats (zeroeval))

Agents

Toolathlon49.7%SR

Code

SWE Atlas - Codebase QnA46.2%SR
DeepSWE 1.140.4%SR

Reasoning

SWE-bench Multilingual78.5%SR
Terminal-Bench 2.170.2%SR
SWE-Bench ProPrinceton NLP (2024)59.4%SR

AA Evaluation Indices

(Artificial Analysis)

No AA evaluation data available

LLM Stats Category Scores

(LLM Stats (zeroeval))
Reasoning
60
General
60
Code
60
Tool Calling
60
Agents
50

Pricing

Input Price$0 / 1M tokens
Output Price$0 / 1M tokens
Blended Price (3:1)$0 / 1M tokens
Cache Read PriceFree
Cache Write PriceFree

Speed

No speed data available

Provider Price Ranking

Provider Price Ranking

6 providers

Cheapest: PoolsideMost Expensive: Pioneer
ProviderInputOutput
1PoolsidePRIMARY
$0
$0
2OpenRouter
$0.09
$0.18
3NanoGPT
$0.1
$0.2
4Kilo Gateway
$0.1
$0.2
5Vercel AI Gateway
$0.1
$0.2
6Pioneer
$0.1
$0.2

Compare pricing across different API providers for this model.

External Sources