Poolside Releases Laguna S 2.1: 118B MoE Model Runs on a Single DGX
joecole · x · 2026-07-22
Poolside has released Laguna S 2.1, a highly efficient model designed for local inference. It features a total of 118B parameters but utilizes a Mixture-of-Experts (MoE) architecture, activating only 8B parameters per token while supporting a 1M token context window. This allows it to outperform models ten times its size, like DeepSeek V4 Pro Max, on the Terminal-Bench while running entirely on a single NVIDIA DGX Spark. Weights are available on Hugging Face under the OpenMDW-1.1 license.
Related event: Poolside Releases 118B Open-Weight Coding Model Laguna S 2.1(35 posts)→
More from Infra
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11
- PlanetScale launches sharded Postgres: 768 servers acting as one, 1PB scale — dhruv2038 · 2026-09-11