Improved#295
OpenAI runs show what they cost
Runs on OpenAI’s models now show their cost on the History page, next to Claude’s, and count toward LLM spend. MergeLoop prices both from each vendor’s list prices per token, including cached tokens and cache writes.
Runs from before today keep showing no cost. Runs on a ChatGPT plan show none either, since the plan bills no tokens. A run that sends a very long prompt can cost more than shown, because vendors charge more above a prompt length that MergeLoop cannot see per request.