It’s more that they have a different business case than competing for the top spots on public benchmarks.

They seem to be oriented more toward customizing models for the concrete needs of a company, on-prem deployment, proprietary knowledge-bases, etc.

It has been more competitive than any other LLMs by non-US/CN companies. Including Korean and Japanese models in case you want to bring those up.

It has produced some locally useful models.