Executive Summary
Poolside has launched its Laguna S 2.1 model, a specialized coding AI, on the AI Gateway platform. This open-weight Mixture-of-Experts (MoE) model features a 1 million token context window and is designed for complex, long-running agentic tasks like code generation, debugging, and MLOps. The model is offered in both free and paid tiers, targeting developers and AI researchers who require advanced coding capabilities through a unified API.
Key Takeaways
* Product Name: Laguna S 2.1
* Primary Function: An open-weight MoE model specializing in agentic coding tasks.
* Key Capabilities:
* Supports a 1 million token context window.
* Operates in "thinking" and "no-thinking" modes.
* Excels at writing/debugging code, running tests, and working on MLOps pipelines.
* Reported high performance on coding benchmarks like Terminal-Bench 2.1 (70.2%) and SWE-bench Multilingual (78.5%).
* Availability: Immediately available on AI Gateway.
* Pricing Tiers:
* Free: `poolside/laguna-s-2.1-free`
* Paid: `poolside/laguna-s-2.1` (offers higher throughput and rate limits).
Strategic Importance
This launch makes a powerful, specialized coding model with a very large context window easily accessible to developers, strengthening AI Gateway's position as a versatile model router.