The right model for the right prompt.
Every time.

or try

0.878 correlation with real model performance · measured across 12 model families

How complex is this prompt?

Cognitive load
00 /100     
Difficulty breakdown

Summary and drivers appear after analysis.

▶ LLM STRESS TEST ◂ AWAITING ◂ -- LAYERS ●STBY
PEAK -- · SETTLE -- · ΔH --
IDLE
00%
LAYER=--/-- · DEPTH=-- · H[?]=-- · PHASE=IDLE

Which model should you use?

Recommended route
Cheapest that works →         

Reasoning budget