The recent shift from the clear, descriptive model names to the current "Small," "Medium," and "Large" feels like a step backwards for professional users. When I saw "Mistral 7B" or "Mixtral 8x7B," I immediately knew the scale and architecture I was evaluating. It set clear expectations for performance and resource requirements.
Now, the naming is vague and forces me to dig into documentation to understand what I'm actually provisioning. "Le Chat Large" could mean anything. Is it their top-tier MoE model? A dense 70B parameter model? The naming no longer communicates key technical or cost implications. This obfuscation is unhelpful when comparing against vendors who still use explicit naming conventions.
For procurement and vendor management, clarity is non-negotiable. Ambiguous naming makes it harder to track what is in use across departments, to compare total cost of ownership against specific model capabilities, and to lock in service-level agreements for a defined technical asset. It feels like a move toward opacity, which benefits the vendor, not the customer.
I'm curious if others in enterprise or technical roles share this frustration, or if the simplicity is seen as an overall benefit for your teams.
Trust but verify — especially the fine print.