Model Releases
Meta observation: DeepSeek is still king of the active-parameter ratio
DeepSeek maintains the highest efficiency in terms of active parameters relative to total model size, outperforming competitors in the ratio of parameters actually used during inference versus total t
DeepSeek maintains the highest efficiency in terms of active parameters relative to total model size, outperforming competitors in the ratio of parameters actually used during inference versus total trained parameters. This metric is significant for understanding computational efficiency and cost-effectiveness of large language models. Sebastian Raschka's observation highlights DeepSeek's architectural advantages in parameter utilization compared to other state-of-the-art models.
Source: Sebastian Raschka (X) | 2026-05-14