Make model pricing badges more granular (5-point Logarithmic Scale)
Right now, the 4 pricing buckets ($, $$, $$$, and red $$$) end up a bit too compressed because API pricing varies exponentially. For example, two models might both show as $$, but one is actually a fraction of the cost of the other.
Since keeping the simple green/red color scheme is great for UI cleanliness and avoids the contrast issues of yellow or orange, it would be awesome to just expand the scale to 5 states and map them logarithmically:
$(Green) — Ultra-economy$$(Green) — Cheap$$$(Green) — Mid-range$$$$(Red) — Premium / Expensive$$$$$(Red) — Ultra-expensive outliers
This keeps the clear mental model (Green = economy, Red = expensive warning) but gives enough granularity so models with vastly different prices (like Gemini Pro vs. GLM-5) don't get forced into the exact same bucket.
Log in to comment and vote
Comments1
Mark Florkowski
There actually already are 5 states in the current model:
… (Grey) — Ultra-economy
$(Green) — Cheap$$(Green) — Mid-range$$$(Green) — Premium / Expensive$$$+(Red) — Ultra-expensive outliersThe indicators are based on both the per-token cost, as well as the “real world” cost based on token usage (if a “cheap” model uses a lot of tokens to respond, it can be ranked as more expensive than one with a higher per-token cost that is more token efficient)
They aren’t?