Speech is Cheap
Benchmarks
0 versions
Loading live data
Version: None
Features: Any
Format: Any
Time by input
Time to transcribe media by their durations
Average response time in seconds. Lower is better. E2E includes inference provider delay and execution.
No benchmark rows match the current filters.
Metrics
E2E
- or -
Queue Delay
Inf. provider execution
- or -
Decode
VAD
Classify
Diarize
Langify
Overlap
Transcribe
TTFT
Input Bucket Summary
Feature Splits