Laguna S 2.1
Model ● Covered in 4 stories + Follow
Laguna S 2.1 is a 118-billion-parameter mixture-of-experts model released by Poolside AI with open weights on Hugging Face, designed for coding and agentic tasks while activating only 8 billion parameters per token. According to benchmark results reported in recent coverage, the model achieved 78.5% on SWE-Bench Multilingual and 70.2% on Terminal-Bench 2.1, reportedly outperforming larger competitors like DeepSeek-V4-Pro-Max while requiring lower computational cost and fitting on a single high-end GPU at 4-bit quantization.
Updated 7 August 2026
Specifications
No specifications recorded yet.
Latest developments
[AINews] "Laguna S 2.1 Released: Cheaper than Deepseek v4 Flash, Better than V4 Pro"
Latent Space · 1 month ago ·
28
Inside the Model Factory — Eiso Kant, Poolside AI
Latent Space · 1 month ago ·
8
Introducing Laguna S 2.1
Poolside · 1 month ago ·
14
Poolside Releases Laguna S 2.1, an Open-Weight Agentic Coding Model Punching Above Its Weight Class on SWE-Bench Multilingual
MarkTechPost · 1 month ago ·
6
July 2026
Poolside AI releases Laguna S 2.1, an open-weight 118B-parameter coding model with sparse mixture-of-experts architecture Open source release
Relationships
Products & technology
- Poolside AI develops this model · 3 sources
- Poolside develops this model · 1 source