Cost vs Performance
Overall score (higher is better) vs total run cost (lower is better). Green quadrant is the target.
Pure Cost
Total run cost per model (lower is better).
Total Tokens vs Performance
Generated tokens (lower is better) vs score (higher is better).
Total Tokens
Total generated tokens per model.
Model Time vs Performance
Selected score (higher is better) vs total model time (lower is better).
Total Model Times
Total time taken by each model across all requests.