StepFun: APIプロバイダーのベンチマークと分析
Artificial Analysis単語誤り率インデックス、速度、料金を含むパフォーマンス指標でのStepFun APIプロバイダーの分析。
注目情報
API別のArtificial Analysis単語誤り率（AA-WER）インデックス
API別のArtificial Analysis単語誤り率（AA-WER）インデックス
% of words transcribed incorrectly · Lower is better · AA-WER v2 incorporates 3 datasets: AA-AgentTalk (50%), VoxPopuli-Cleaned-AA (25%), Earnings22-Cleaned-AA (25%)
注: Earnings22で、時間制限によりモデルがフル尺の音声を安定して処理できない場合、約9分ごとに分割します（対象: Nova 2 Pro, Amazon GPT-4o Transcribe, OpenAI GPT-4o Mini Transcribe, OpenAI）。さらに短い時間制限があるモデルでは、約30秒ごとに分割します（対象: Canary Qwen 2.5B, NVIDIA）。
APIベンチマーク
Artificial Analysis単語誤り率インデックス vs. 料金
% 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
Most attractive quadrant
速度係数
速度係数
Input audio seconds transcribed per second · Higher is better
料金
文字起こし料金
USD per 1000 minutes of audio
主要指標の概要と詳細情報
プロバイダー | 詳細情報 | ||||
|---|---|---|---|---|---|
Qwen3.5 Omni Flash |  Alibaba Cloud | 13.5% | 78.4 | 0.00 | |
Qwen3.5 Omni Plus | Alibaba Cloud | 3.5% | 94.5 | 0.00 | |
Nova 2 Pro |  Amazon Bedrock | 4.9% | 23.0 | 3.10 | |
Amazon Transcribe | Amazon Bedrock | 4.1% | 15.3 | 6.00 | |
Universal-3 Pro |  AssemblyAI | 3.1% | 110.4 | 3.50 | |
Universal, AssemblyAI | AssemblyAI | 3.8% | 123.5 | 2.50 | |
MAI-Transcribe-1.5 |  Microsoft Azure | 2.4% | 190.3 | 6.00 | |
MAI-Transcribe-1 | Microsoft Azure | 2.6% | 66.6 | 6.00 | |
transcribe-03-2026 |  Cohere | 4.6% | 118.7 | 0.00 | |
Nova-3 |  Deepgram | 5.2% | 545.7 | 4.30 | |
Scribe v2 |  ElevenLabs | 2.2% | 57.0 | 3.67 | |
Solaria-1, Gladia |  Gladia | 4.1% | 79.8 | 10.17 | |
Solaria-3, Gladia | Gladia | 3.2% | 62.6 | 10.16 | |
Gemini 3.1 Pro Preview (High) |  Google | 2.8% | 7.0 | 18.15 | |
Gemini 3.1 Pro Preview (Low) | Google | 3.6% | 6.8 | 7.72 | |
Gemini 3 Flash (High) | Google | 2.9% | 16.6 | 13.70 | |
Gemini 2.5 Flash Lite | Google | 5.2% | 72.5 | 6.56 | |
Gemini 2.5 Flash | Google | 5.1% | 75.2 | 6.66 | |
Gemini 2.5 Pro | Google | 2.9% | 13.2 | 11.39 | |
Gemini 3.1 Flash-Lite Preview (Minimal) | Google | 3.4% | 80.5 | 5.83 | |
Gradium Speech-to-Text |  Gradium | 6.8% | 2.3 | 13.00 | |
Grok Speech to Text, SpaceXAI |  SpaceXAI | 4.0% | 230.1 | 1.67 | |
Voxtral Mini Transcribe 2 |  Mistral | 3.6% | 84.5 | 3.00 | |
Voxtral Small | Mistral | 2.8% | 64.4 | 4.00 | |
Voxtral Mini |  DeepInfra | 3.8% | 78.9 | 1.00 | |
Modulate STT Batch English VFast |  Modulate | 4.2% | 62.0 | 0.42 | |
Parakeet TDT 0.6B V3, Togetherai |  Together AI | 4.5% | 280.4 | 1.50 | |
Canary Qwen 2.5B, NVIDIA |  Replicate | 4.3% | 5.7 | 0.74 | |
Parakeet TDT 0.6B V2, NVIDIA |  NVIDIA | 6.4% | 99.2 | 0.00 | |
Parakeet RNNT 1.1B | Replicate | 5.4% | 6.3 | 1.91 | |
GPT Transcribe, OpenAI |  OpenAI | 3.3% | 37.6 | 4.50 | |
GPT-4o Transcribe | OpenAI | 4.0% | 37.8 | 6.00 | |
GPT-4o Mini Transcribe | OpenAI | 4.5% | 43.2 | 3.00 | |
Smallest AI Pulse Pro |  Smallest.ai | 2.4% | 275.9 | 4.00 | |
Resonant-1 |  Reson8 | 3.4% | 325.5 | 3.60 | |
Rev AI |  Rev AI | 5.9% | 12.8 | 3.33 | |
Smallest AI Pulse | Smallest.ai | 4.4% | 283.2 | 5.00 | |
Soniox v5 Async |  Soniox | 3.8% | 23.5 | 1.66 | |
Soniox V4 | Soniox | 3.9% | 28.5 | 1.66 | |
Speechmatics Melia |  Speechmatics | 4.9% | 207.5 | 4.00 | |
Speechmatics Standard | Speechmatics | 5.1% | 106.3 | 7.50 | |
Speechmatics Enhanced | Speechmatics | 4.0% | 74.2 | 12.50 | |
Whisper Large v3 Turbo |  Groq | 4.6% | 151.1 | 0.67 | |
Wizper Large v3 |  fal.ai | 4.7% | 174.9 | 0.50 | |
Incredibly Fast Whisper | Replicate | 5.7% | 54.1 | 1.49 | |
Whisper Large v3 | Replicate | 10.1% | 2.6 | 4.23 | |
Whisper Large v3 | fal.ai | 4.1% | 85.2 | 1.15 | |
Whisper Large v3 | Together AI | 4.5% | 287.9 | 1.50 | |
Whisper Large v2 | OpenAI | 4.1% | 28.9 | 6.00 |