Poolside AI releases Lagona S2.1, a 118B MoE coding model runnable on consumer hardware

Poolside AI has released Lagona S2.1, an 118-billion-parameter mixture-of-experts model designed for local deployment by developers. Despite its large parameter count, only a fraction are active per token, allowing it to run on consumer hardware like an RTX 3090 or a MacBook. The model demonstrates strong performance on coding benchmarks, outperforming models like Gemini 3.5 Flash and Claude Sonnet, and is capable of generating complex applications and simulations. AI

IMPACT Enables developers to run powerful coding models locally, potentially increasing productivity and reducing reliance on cloud infrastructure.

RANK_REASON New model release from a mid-size AI lab with notable efficiency and local deployment capabilities. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

• Claude Fable 5
• Claude Sonnet
• Gemini 3.5 Flash
• GLM 5.2
• GPT-5.6
• Kimi K2.7 Code
• Kimi K3
• Lagona S2.1
• MacBook
• Nvidia DGX Spark
• Poolside AI
• Qwen 3.6
• RTX 3090

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →