Zyphra 發佈 Zamba2-VL 視覺語言模型系列，提供 1.2B、2.7B 和 7B 參數規模

ModelZamba2-VL: Hybrid SSM Vision-Language ModelsZyphra is releasing Zamba2-VL, a family of open vision-language models built on the Zamba2 hybrid SSM–Transformer backbone available at 1.2B, 2.7B, and 7B parameters. The models are competitive in accuracy with leading open Transformer-based vision-language models at comparable scales while running substantially faster.June 2, 2026

來源：Zyphra Research · zyphra.com