B
Prompt Benchmark
Compare raw vs compiled output
Raw = your prompt as-is · Compiled = Prompt Compiler's polished version
Cheap and surprisingly capable. Good default for everyday tests.
[cheap] runs fast and costs little. [balanced] is smarter but slower. [preview] is a newer model still being tested.
Ready
Enter the prompt you want to benchmark across different models.
B
No benchmark yet
Paste a prompt on the left, tweak the OpenRouter model if needed, then run a benchmark.