It seems weird that a likely prime indicator of capability isn't mentioned, the model size.
I thought most frontier LLM providers don’t disclose exact parameter counts these days.
I thought most frontier LLM providers don’t disclose exact parameter counts these days.