News

Sofya deployed Llama models with knowledge distillation techniques, reducing provider documentation time by 30% while maintaining 90% customer satisfaction.

The marketplace debuts with 200+ partners offering pre-built AI components, addressing a key bottleneck in enterprise AI adoption—time and expertise to build capabilities.

The system automates converting meeting transcripts to PRDs and development tickets using specialized agents powered by Mistral Large 2, eliminating manual effort.

Med-Gemini, a multimodal model fine-tuned on de-identified medical data, achieves 91.1% accuracy on medical licensing exams and interprets 3D scans.

Anthropic's Claude 3.7 Sonnet sets new benchmarks in coding abilities. Novo Nordisk uses it to reduce clinical report writing from 12 weeks to 10 minutes.

Claude 3.7 Sonnet, the first hybrid reasoning model, will be tested by scientists to compress decades of scientific progress into shorter timeframes.