Dev Tools · 20h ago
Poolside's Laguna S 2.1 open-weight model debuts on Vercel AI Gateway
Vercel's AI Gateway now hosts Poolside's Laguna S 2.1, an open-weight Mixture-of-Experts model with up to 1M token context and thinking/no-thinking modes. It achieves 70.2% on Terminal-Bench 2.1 and 78.5% on SWE-bench Multilingual in thinking mode. The model specializes in agentic coding tasks like debugging, testing, and MLOps.
Meridian48 take
While the benchmark scores are solid, the real test is whether developers will adopt an open-weight model over proprietary alternatives for production agentic coding workflows.
ai-modelsdeveloper-tools