Poolside drops Laguna S 2.1, an open-weight coding model that beats rivals 10x its size

Poolside, a San Francisco-based AI lab, unveiled its most advanced coding model, Laguna S 2.1, which surprisingly outperforms much larger competitors despite having just 8 billion active parameters. This model challenges the conventional wisdom that size matters in AI, advocating instead for radical transparency and efficiency. By releasing Laguna S 2.1 openly, Poolside is making a bold statement about the future of AI development, suggesting that smaller labs can compete effectively by prioritizing innovation and transparency over sheer scale.