THE EVIDENCE / VALUE
Value
benchmarks.
Performance relative to cited API input and output prices. Results are grouped by benchmark and unit. Different tests are never merged into a raw leaderboard.
No sourced records yet. This panel will populate after verified CSV data is imported.
Limitations: Provider discounts and cached-token pricing may differ from list prices. Publication methods, prompts, sampling and model versions may differ. Check each cited source before interpreting a result.