The question is not, will mistral be able to create the best models (seems bloody unlikely). The question is, will mistral have enough expertise to dominate its business use-case, which is the model infra/deployment market within the EU (which is more plausible, with even a 3rd, 4th, 5th best model development team).