Laguna S 2.1
Laguna S 2.1 is a model for agentic coding and long-horizon work. It has 118B total parameters, 8B active parameters per token, and supports up to 1M tokens of context.See Model release notes.July 2026 inference deployment bundle
The July 2026 inference deployment bundle supports cloud and on-premises model inference deployments. Use it to deploy Poolside Inference with the latest inference Helm chart, theatlas inference server image, SLES 16 support, and pool CLI release v1.0.8.See Poolside Inference release notes.Laguna XS 2.1
Laguna XS 2.1 is an incremental update to Laguna XS.2 that adds native reasoning support and improves performance on multilingual coding and terminal-style tasks. It supports a 256K context window.See Model release notes.Laguna M.1
Laguna M.1 is available for agentic coding workflows that require multiple steps, tool use, and validation. It supports a 256K context window and improves Poolside’s reference benchmark results compared with Malibu 2.2.See Model release notes.Laguna XS.2
Laguna XS.2 is available for faster agentic coding workflows, especially when you need strong coding performance on hardware with GPU constraints. It supports a 256K context window.See Model release notes.Malibu 2.2
Malibu 2.2 supports agentic coding workflows that require tool use across multiple turns. Use it for direct coding tasks where the model can gather context, edit files, and run validation commands.See Model release notes.Malibu 2.1
Malibu 2.1 improves code editing, instruction following,.poolside instruction support, long-context performance, and agent tool calling compared with Malibu 2.0.See Model release notes.