DeepSeek launches V4.1-Flash with faster inference and higher throughput

The Chinese startup said the latest model focuses on greater capability, faster inference, higher throughput and scaling to larger models