High-performance inference framework for large language models, focusing on efficiency, flexibility, and availability.
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).
