Poolside Releases Laguna S 2.1, an Open-Weight Agentic Coding Model Punching Above Its Weight Class on SWE-Bench Multilingual

Poolside released Laguna S 2.1, a 118B-parameter open-weight agentic coding model with about 8B active parameters per token.

The model has a 1M-token context window, multiple published weight formats, and was scaled up from Laguna XS 2.1 in under nine weeks.

Published results show strong performance on coding benchmarks, including a leading score on SWE-Bench Multilingual versus several much larger models.

Poolside says Laguna S 2.1 can run on smaller hardware, aiming to make capable coding agents cheaper to deploy.