Run Benchmark Replay

Flagship free live smoke

No API key required. Live baseline vs Omar/RCC run. Smoke test only, not benchmark proof.

For serious verification, run the 100-sample BYOK board-depth reproduction.

Checking free smoke availability.

Preparing free smoke 0%

Ready. Click the free smoke button to start a 20-sample live run.

Select one benchmark lane and run current live baseline vs Omar/RCC outputs.

API key

This run uses your key directly for the selected benchmark samples. Your key is not stored and is cleared after the run.

Model Sample count Benchmark subset

One current live lane per benchmark. Historical proof and diagnostic paths stay out of the primary selector.

Loading sample counts.

Loading access state.

Developer/debug details

Ready.