Speech to Text (ASR) Providers Leaderboard & Comparison | Artificial Analysis
K
Speech to Text AI Model & Provider Leaderboard
Compare word error rate, speed, and pricing across Speech to Text models and providers.
For further details, see our methodology page.
Related Links
Highlights
WER Index (Non-streaming)
AA-WER v2 · % of words transcribed incorrectly · Lower is better
Speed Factor
Input audio seconds transcribed per second · Higher is better
Price
USD per 1000 minutes of audio · Lower is better
Artificial Analysis Word Error Rate Index (Non-streaming)AA-WER (Non-streaming) by DatasetCleaned Dataset ComparisonAPI Benchmarks
Artificial Analysis Word Error Rate Index (Non-streaming)
Artificial Analysis Word Error Rate Index (Non-streaming)
% of words transcribed incorrectly · Lower is better · AA-WER v2 incorporates 3 datasets: AA-AgentTalk (50%), VoxPopuli-Cleaned-AA (25%), Earnings22-Cleaned-AA (25%)
AllProprietaryOpen Weights
30 of 64 models
Note: For Earnings22, if a model cannot reliably handle full-length audio due to time limits, we chunk to ~9 minutes (relevant to: GPT-4o Mini Transcribe, OpenAI; GPT-4o Transcribe, OpenAI; Nova 2 Pro, Amazon; Voxtral Mini Transcribe, Mistral). For models with even shorter time limits, we chunk to ~30 seconds (relevant to: Qwen3 ASR Flash, Alibaba; Parakeet TDT 0.6B V3, NVIDIA; Canary Qwen 2.5B, NVIDIA).
Artificial Analysis Word Error Rate (AA-WER) Index
AA-WER (Non-streaming) by Dataset
AA-WER (Non-streaming): AA-AgentTalk Dataset
% of words transcribed incorrectly on the AA-AgentTalk dataset · Lower is better
30 of 64 models
Artificial Analysis Word Error Rate (AA-WER) Index
Cleaned Dataset Comparison
VoxPopuli: Cleaned vs Original Subset of Publicly Available Data
% WER (word error rate) · Lower is better
Sort by
CleanedOriginal
30 of 32 models
VoxPopuli-Cleaned-AAVoxPopuli
Note: The cleaned versions remove transcription errors from the reference text, providing a more accurate ground truth for model evaluation.
Artificial Analysis Word Error Rate (AA-WER) Index
API Benchmarks
Artificial Analysis Word Error Rate Index (Non-streaming) vs. Price
% of words transcribed incorrectly · Lower is better · AA-WER v2 incorporates 3 datasets: AA-AgentTalk (50%), VoxPopuli-Cleaned-AA (25%), Earnings22-Cleaned-AA (25%) · USD per 1000 minutes of audio
30 of 50 models
Most attractive quadrant
GladiaMicrosoft AzurexAIGoogleSmallest.aiMistralAssemblyAIElevenLabsAmazon BedrockTogether.aiReplicateOpenAIDeepgramfal.aiFireworksSpeechmaticsRev AIModulate
Artificial Analysis Word Error Rate (AA-WER) Index
Price
Speed Factor
Input audio seconds transcribed per second · Higher is better
30 of 50 models
Speed Factor
Price of Transcription
USD per 1000 minutes of audio
30 of 50 models
Price
Summary of Key Metrics & Further Information
Model| Provider| Whisper Version| Word Error Rate%| Median Speed Factor| PriceUSD per 1,000 minutes| Further Details
Qwen3.5 Omni Flash| Alibaba Cloud| 13.5%| 73.8| 0.00| Details Qwen3.5 Omni Plus| Alibaba Cloud| 3.5%| 99.0| 0.00| Details Nova 2 Pro| Amazon Bedrock| 4.9%| 22.4| 3.10| Details Amazon Transcribe| Amazon Bedrock| 4.1%| 19.0| 24.00| Details Universal-3 Pro| AssemblyAI| 3.1%| 83.9| 3.50| Details Universal, AssemblyAI| AssemblyAI| 3.8%| 113.3| 2.50| Details MAI-Transcribe-1.5| Microsoft Azure| 2.4%| 174.1| 6.00| Details MAI-Transcribe-1| Microsoft Azure| 2.6%| 56.4| 6.00| Details transcribe-03-2026| Cohere| 4.6%| 64.2| 0.00| Details Nova-3| Deepgram| 5.2%| 510.5| 4.30| Details Nova-2| Deepgram| 5.3%| 483.1| 4.30| Details Base| Deepgram| 10.7%| 347.3| 12.50| Details Scribe v2| ElevenLabs| 2.2%| 34.5| 3.67| Details Scribe v1| ElevenLabs| 3.0%| 39.3| 6.67| Details Solaria-1, Gladia| Gladia| 4.1%| 60.9| 4.07| Details Solaria-3, Gladia| Gladia| 3.2%| 60.1| 10.16| Details Gemini 3.1 Pro Preview (High)| Google| 2.8%| 6.7| 18.15| Details Gemini 3.1 Pro Preview (Low)| Google| 3.6%| 7.0| 7.72| Details Gemini 3 Flash (High)| Google| 2.9%| 16.9| 13.70| Details Gemini 2.5 Flash Lite| Google| 5.2%| 69.3| 6.56| Details Gemini 2.5 Flash| Google| 5.1%| 70.5| 6.66| Details Gemini 2.5 Pro| Google| 2.9%| 12.2| 11.39| Details Gemini 3.1 Flash-Lite Preview (Minimal)| Google| 3.4%| 75.5| 5.83| Details Gradium Speech-to-Text| Gradium| 8.4%| 2.3| 13.00| Details Grok Speech to Text, xAI| xAI| 4.0%| 100.5| 1.67| Details LLM Speech, Azure| Microsoft Azure| 3.7%| 57.5| 6.00| Details Voxtral Mini Transcribe 2| Mistral| 3.6%| 74.1| 3.00| Details Voxtral Mini Transcribe| Mistral| 3.5%| 70.9| 2.00| Details Voxtral Small| Mistral| 2.8%| 66.1| 4.00| Details Voxtral Mini| DeepInfra| 3.8%| 74.2| 1.00| Details Modulate STT Batch English VFast| Modulate| 6.3%| 188.9| 0.42| Details Parakeet TDT 0.6B V3, Togetherai| Together.ai| 4.5%| 995.4| 1.50| Details Canary Qwen 2.5B, NVIDIA| Replicate| 4.3%| 5.9| 0.74| Details Parakeet TDT 0.6B V2, NVIDIA| NVIDIA| 6.4%| 100.9| 0.00| Details Parakeet RNNT 1.1B| Replicate| 5.4%| 6.3| 1.91| Details GPT-4o Transcribe| OpenAI| 4.0%| 33.6| 6.00| Details GPT-4o Mini Transcribe| OpenAI| 4.5%| 48.2| 3.00| Details Rev AI| Rev AI| 5.9%| 12.9| 3.33| Details Smallest AI Pulse| Smallest.ai| 4.4%| 135.3| 5.00| Details Speechmatics Standard| Speechmatics| 5.1%| 68.2| 4.00| Details Speechmatics Enhanced| Speechmatics| 4.0%| 54.1| 6.70| Details Whisper Large v3 Turbo| Groq| v3 Turbo| 4.6%| 119.7| 0.67| Details Whisper Large v3 Turbo| Fireworks| v3 Turbo| 4.7%| 140.8| 1.00| Details Wizper Large v3| fal.ai| large-v3| 4.7%| 194.6| 0.50| Details Incredibly Fast Whisper| Replicate| large-v3| 5.7%| 53.6| 1.49| Details Whisper Large v3| Replicate| large-v3| 10.1%| 2.7| 4.23| Details Whisper Large v3| fal.ai| large-v3| 4.1%| 98.7| 1.15| Details Whisper Large v3| Fireworks| large-v3| 4.6%| 313.7| 1.00| Details Whisper Large v3| Together.ai| large-v3| 4.5%| 415.8| 1.50| Details Whisper Large v2| OpenAI| large-v2| 4.1%| 26.7| 6.00| Details
Frequently Asked Questions
Which is the most accurate speech to text model?
What are the top speech to text models?
Which is the fastest speech to text model?
Which is the cheapest speech to text model?
Which is the best open weights speech to text model?
What are the top open weights speech to text models?
How do I choose the best speech to text model?
Speech to Text models & providers compared: Solaria-3, MAI-Transcribe-1.5, MAI-Transcribe-1, Grok Speech to Text, xAI, Qwen3.5 Omni Flash, Qwen3.5 Omni Plus, Cohere Transcribe 03-2026, Gemini 3.1 Pro Preview (High), Gemini 3.1 Pro Preview (Low), Smallest AI Pulse, Voxtral Mini Transcribe 2, Universal-3 Pro, Scribe v2, Gemini 3 Flash (High), Nova 2 Pro, Gradium Speech-to-Text, LLM Speech, Azure, Parakeet TDT 0.6B V3, Togetherai, Gemini 2.5 Flash Lite, Canary Qwen 2.5B, Replicate, Voxtral Mini Transcribe, Voxtral Small, Voxtral Mini, Deepinfra, Gemini 2.5 Flash, Gemini 2.5 Pro, Parakeet TDT 0.6B V2, Solaria-1, GPT-4o Transcribe, GPT-4o Mini Transcribe, Scribe v1, Nova-3, Universal, Whisper (L, v3, Turbo), Groq, Whisper (L, v3, Turbo), Fireworks, Parakeet RNNT 1.1B, Replicate, Amazon Transcribe, Wizper (L, v3), fal.ai, Incredibly Fast, Replicate, Whisper (L, v3), Replicate, Whisper (L, v3), fal.ai, Whisper (L, v3), Fireworks, Whisper Large v3, together.ai, Nova-2, Standard, Enhanced, Large v2, Base, Rev AI, Gemini 3.1 Flash-Lite Preview (Minimal), Modulate STT Batch English VFast.