Dataset · Speech recognition
Speech Recognition Speed: Real-Time Factor Data
How much faster than real time do speech recognition models transcribe audio, and on which hardware?
24 data points from 0 sourcesUpdated October 19, 2026Free to reuse under CC BY 4.0
Every data point
Each row is one published figure for one group, in times faster than real time. The row opens its fact page with the exact words of the source; the source opens the original.
| Source | |||||||
|---|---|---|---|---|---|---|---|
| Parakeet TDT v3, Spanish (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 22 | Calculated |
| Parakeet TDT v3, Portuguese (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 22 | Calculated |
| Parakeet TDT v3, Russian (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 21 | Calculated |
| Parakeet TDT v3, Italian (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 21 | Calculated |
| Parakeet TDT v3, German (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 20 | Calculated |
| Parakeet TDT v3, Slovenian (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 20 | Calculated |
| Parakeet TDT v3, English (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 19 | Calculated |
| Parakeet TDT v3, French (M2 Pro) | Parakeet TDT 0.6B v3 | Open model | Apple M2 Pro | FLEURS | 2026 | 18 | Calculated |
| Whisper large-v3-turbo, Italian (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 17 | Calculated |
| Whisper large-v3-turbo, Portuguese (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 16 | Calculated |
| Whisper large-v3-turbo, German (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 16 | Calculated |
| Whisper large-v3-turbo, Spanish (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 15 | Calculated |
| Whisper large-v3-turbo, Latvian (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 15 | Calculated |
| Whisper large-v3-turbo, Russian (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 15 | Calculated |
| Whisper large-v3-turbo, Ukrainian (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 13 | Calculated |
| Whisper large-v3-turbo, English (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 12 | Calculated |
| Whisper large-v3-turbo, French (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 12 | Calculated |
| Whisper large-v3-turbo, Slovenian (M2 Pro) | Whisper large-v3-turbo | Open model | Apple M2 Pro | FLEURS | 2026 | 12 | Calculated |
| Apple SpeechTranscriber, Portuguese (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 51,177 | Calculated |
| Apple SpeechTranscriber, Italian (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 75 | Calculated |
| Apple SpeechTranscriber, German (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 72 | Calculated |
| Apple SpeechTranscriber, Spanish (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 57 | Calculated |
| Apple SpeechTranscriber, French (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 46 | Calculated |
| Apple SpeechTranscriber, English (M2 Pro) | Apple SpeechTranscriber | Built-in dictation | Apple M2 Pro | FLEURS | 2026 | 34 | Calculated |
24 rows. Open a row for the exact words of its source.
Calculated from the rows
Every statistic is calculated by code from the rows above. Open one to see the exact values it summarises.
How this dataset is built
- What counts as a row
- Processing speeds of speech recognition systems expressed as an inverse real-time factor (seconds of audio per second of compute) or convertible to one (a real-time factor of 0.05 is 20 times real time), with the hardware stated.
- What does not
- Latency to the first word (a different measure) and speeds without the hardware.
- How rows are found and checked
- Rows come from the facts already cited across the library, from public leaderboards, and from model papers. A real-time factor is converted to its inverse and the conversion noted on the row.
- What the data cannot tell you
- Speed depends on hardware, batch size and audio length. Compare rows measured on the same hardware.
The search record
157
works found by the searches below
55
report a figure in the dataset's terms
0
read and judged by a researcher
0
kept, with their figures checked
The 6 searches, last run October 19, 2026
- OpenAlex: “real-time factor speech recognition model inference speed”, 50 results
- arXiv: “real-time factor speech recognition model inference speed”, 23 results
- OpenAlex: “inverse real-time factor ASR leaderboard”, 50 results
- arXiv: “inverse real-time factor ASR leaderboard”, 2 results
- OpenAlex: “on-device speech recognition latency real-time factor”, 50 results
- arXiv: “on-device speech recognition latency real-time factor”, 7 results
The whole method, from search to calculation, is on the research data page.
Cite and reuse
You may copy, share and adapt this dataset for any purpose, including commercially, if you credit it with a link. Quoted passages remain their authors' words: cite the original source for those.
Yap (2026). Speech recognition processing speed [Data set]. https://yapapp.cc/data/asr-speed
License: CC BY 4.0.
More datasets
- On-Device Speech Recognition Benchmark on a Mac26 data points
- Typing Speed Statistics: Average WPM by Group14 data points
- Speech vs Typing Speed: Text Entry Rates Compared57 data points
- Speaking Rate Statistics: Words per Minute7 data points
- Speech Recognition Accuracy: Word Error Rates597 data points
- How Many Languages Speech Recognition Supports45 data points
- Speech-to-Text API Pricing: Cost per Hour of Audio34 data points
- Dictation Software Prices: What Each App Costs37 data points
- How Long Voice Assistants Keep Your Recordings38 data points
- Repetitive Strain Injury Statistics: How Common Is RSI18 data points
- Disability Statistics: Conditions That Make Typing Hard22 data points
- Voice Assistant and Voice Search Usage Statistics2 data points
Read the overview: Speech Recognition: What It Is and How Well It Works