OVH Releases First In-House Open-Source MoE Model Olala mini

French cloud provider OVH has launched Olala, its first in-house LLM family, debuting with the open-source Olala mini 7A1B, a MoE model with 7B total parameters and only 1B active per layer. It runs on 8GB GPUs at up to 160 tokens per second.

2026-11-17 ~ 2026-11-17 · 2 related posts