Poolside's MoE with 3B active parameters for agentic coding and long tasks on a single machine. Works across multiple files and the terminal.
| SWE-bench Verified | 70.9 | |
|---|---|---|
| SWE-bench Multilingual | 63.1 | |
| Terminal-Bench 2.0 | 37.5 |
Which GPU does Laguna XS 2.1 33B-A3B need?
The entry configuration is 1x RTX5090, which delivers around 48 tokens per second. For more parallel requests and headroom on speed, go with 1x H200.
How much does Laguna XS 2.1 33B-A3B cost?
An hour of Laguna XS 2.1 33B-A3B runs from $1.84 to $8.20 depending on the GPU configuration. Billing is hourly: you pay for server time, not for tokens.
How do I call Laguna XS 2.1 33B-A3B from my code?
Once deployed you get an OpenAI-compatible endpoint: point any OpenAI client at the base_url you receive and pass laguna-xs-2.1 as the model name. Code written against the OpenAI API needs no changes.
Does the data stay private?
Yes. Laguna XS 2.1 33B-A3B runs on a dedicated GPU server inside a private environment: requests and responses never reach the model vendor and are not used for training.
What context window and license does Laguna XS 2.1 33B-A3B have?
Laguna XS 2.1 33B-A3B holds 32K tokens of context per request and ships under the OpenMDW-1.1 license, so you can use it commercially on your own hosting terms.