# transcription model rankings

| \# | model | card | cheapest list | cheapest effective | unit | providers | tokens/week |
| --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | `openai/gpt-4o-mini-transcribe` | 1.25 | 1.25 openai | 1.249 openai |  | 1 | 14638499 |
| 2 | `google/gemini-3.5-transcribe` |  | 2.0 google-ai-studio | 2.0 google-ai-studio |  | 1 | 5347309 |
| 3 | `openai/gpt-4o-transcribe` | 2.5 | 2.5 openai | 2.4988 openai |  | 1 | 1664984 |
| 4 | `mistralai/voxtral-mini-transcribe` | 50.0 | 50.0 mistral | 143.6899 mistral | $0.00005/second Audio Seconds mistral | 1 | 396011 |
| 5 | `mistralai/voxtral-mini-transcribe-3-0` | 64.1667 | 58.3333 mistral/zdr | 5.4421 mistral/zdr | $0.0000583333/second Audio Seconds mistral/zdr | 1 | 360048 |
| 6 | `mistralai/voxtral-small-transcribe-3-0` | 73.3333 | 66.6667 mistral/zdr | 143.5711 mistral/zdr | $0.0000666667/second Audio Seconds mistral/zdr | 1 | 308119 |
| 7 | `qwen/qwen3-asr-flash-2026-02-10` |  | 35.0 alibaba |  | $0.000035/second Audio Seconds alibaba | 1 |  |
| 8 | `x-ai/grok-stt-2.0` | 27.7778 | 27.7778 xai |  | $0.0000277777777778/second Audio Seconds xai | 1 |  |
| 9 | `moondream/parakeet-ultra` |  | 16.6667 primeintellect/bf16 |  | $0.0000166667/second Audio Seconds primeintellect/bf16 | 1 |  |
| 10 | `nvidia/nemotron-3.5-transcribe` |  | 16.6667 primeintellect/bf16 |  | $0.0000166667/second Audio Seconds primeintellect/bf16 | 2 |  |
| 11 | `assemblyai/universal-3-6-pro` | 83.3333 | 83.3333 assemblyai |  | $0.0000833333/second Audio Seconds assemblyai | 1 |  |
| 12 | `elevenlabs/scribe-v2` | 61.1111 | 61.1111 elevenlabs |  | $0.0000611111111111/second Audio Seconds elevenlabs | 1 |  |
| 13 | `elevenlabs/scribe-v2-medical` | 61.1111 | 61.1111 elevenlabs |  | $0.0000611111111111/second Audio Seconds elevenlabs | 1 |  |
| 14 | `fish-audio/transcribe-1-pro` | 100.0 | 100.0 fish-audio |  | $0.0001/second Audio Seconds fish-audio | 1 |  |
| 15 | `assemblyai/universal-3-5-pro` | 83.3333 | 83.3333 assemblyai |  | $0.0000833333/second Audio Seconds assemblyai | 1 |  |
| 16 | `meta/muse-voice-transcribe-1.0` | 50.0 | 50.0 meta |  | $0.00005/second Audio Seconds meta | 1 |  |
| 17 | `microsoft/mai-transcribe-2` |  | 1.0e5 azure |  | $0.1/hour Audio Hours azure | 1 |  |
| 18 | `nvidia/nemotron-3.5-asr-streaming-multilingual-0.6b` |  | 3.33 deepinfra |  | $0.00000333/second Audio Seconds deepinfra/us | 1 |  |
| 19 | `mistralai/voxtral-small-24b-2507-stt` |  | 50.0 deepinfra/bf16 |  | $0.00005/second Audio Seconds deepinfra/bf16 | 1 |  |
| 20 | `mistralai/voxtral-mini-3b-2507` |  | 16.6667 deepinfra/bf16 |  | $0.0000166667/second Audio Seconds deepinfra/bf16 | 1 |  |
| 21 | `qwen/qwen3-asr-1.7b` |  | 7.5 deepinfra/us |  | $0.0000075/second Audio Seconds deepinfra/us | 2 |  |
| 22 | `qwen/qwen3-asr-0.6b` |  | 3.33 deepinfra/us |  | $0.00000333/second Audio Seconds deepinfra/us | 2 |  |
| 23 | `openai/gpt-transcribe` | 75.0 | 75.0 openai |  | $0.000075/second Audio Seconds openai | 1 |  |
| 24 | `fish-audio/transcribe-1` | 100.0 | 100.0 fish-audio |  | $0.0001/second Audio Seconds fish-audio | 1 |  |
| 25 | `x-ai/grok-stt-1.0` | 27.7778 | 27.7778 xai |  | $0.0000277777777778/second Audio Seconds xai | 1 |  |
| 26 | `deepgram/nova-3` | 71.6667 | 71.6667 deepgram |  | $0.0000716666666667/second Audio Seconds deepgram | 1 |  |
| 27 | `microsoft/mai-transcribe-1.5` |  | 3.6e5 azure |  | $0.36/hour Audio Hours azure | 1 |  |
| 28 | `nvidia/parakeet-tdt-0.6b-v3` |  | 16.6667 primeintellect/bf16 |  | $0.0000166667/second Audio Seconds primeintellect/bf16 | 2 |  |
| 29 | `google/chirp-3` |  | 266.6667 google-vertex |  | $0.000266666666667/second Audio Seconds google-vertex | 1 |  |
| 30 | `openai/whisper-large-v3` |  | 7.5 deepinfra |  | $0.0000075/second Audio Seconds deepinfra/us | 3 |  |
| 31 | `openai/whisper-large-v3-turbo` |  | 3.33 deepinfra/us |  | $0.00000333/second Audio Seconds deepinfra/us | 3 |  |
| 32 | `openai/whisper-1` | 100.0 | 100.0 openai |  | $0.0001/second Audio Seconds openai | 1 |  |