Whisper Large v3
OpenAI Whisper Large v3, FP16, 30s English audio, single stream. Higher = faster than real time.
Primary metric: Real-time factor (x RT)
Reference prompts
Representative prompts for this workload. Exact prompts and harness settings still depend on the cited source for each record.
- Prompt 1
30-second clip — clean conference talk, English (LibriSpeech test-clean).
- Prompt 2
30-second clip — podcast crosstalk, 2 speakers, English.
- Prompt 3
30-second clip — accented English (CommonVoice subset).
Reference runtime command
A representative invocation for reproducing this workload class. Source-specific runs may use adjacent runtimes unless the record says otherwise.
whisper-cli -m ggml-large-v3.bin -f sample.wav --language en --output-txtReal-time factor = (audio duration ÷ transcription wall time). Higher = faster than real-time.
Full leaderboard
Every record for Whisper transcription.
No comparable multi-device results
The current selection lacks two devices with matching workload, quantization, context, batch and documented runtime. No speed winner is assigned. Inspect the source-attributed records.
Cite this benchmark
Use this in your paper, blog post, or comparison table.
@misc{myaihardware_whisper-large-v3_2026,
title = {MyAI Bench: Whisper Large v3},
author = {{MyAIHardware Contributors}},
year = {2026},
url = {https://www.myaihardware.com/benchmarks/workload/whisper-large-v3},
note = {Version 1.3, accessed 2026-09-09}
}MyAIHardware Contributors. (2026). MyAI Bench: Whisper Large v3. MyAIHardware. Retrieved 2026-09-09, from https://www.myaihardware.com/benchmarks/workload/whisper-large-v3
MyAIHardware Contributors. "MyAI Bench: Whisper Large v3." MyAIHardware, 2026, https://www.myaihardware.com/benchmarks/workload/whisper-large-v3. Accessed 2026-09-09.
MyAI Bench, Whisper Large v3. MyAIHardware Contributors, 2026. Version 1.3. https://www.myaihardware.com/benchmarks/workload/whisper-large-v3 (accessed 2026-09-09).