Rankings based on ELO scores from blind comparisons by the community.
| Rank | Provider | ELO score | Votes | 95% CI |
|---|---|---|---|---|
| 1 | ElevenLabs (Scribe v2) | 1612 | 325 | ±43 |
| 2 | AssemblyAI (Universal-3 Pro) | 1584 | 325 | ±43 |
| 3 | Gladia (Solaria-3) | 1528 | 325 | ±43 |
| 4 | Mistral (Voxtral Mini) | 1494 | 325 | ±43 |
| 5 | Speechmatics (Enhanced) | 1465 | 325 | ±43 |
| 6 | Deepgram (Nova 3) | 1317 | 325 | ±43 |
Based on 975 blind comparisons. Last updated 2026-10-13.
More comparisons and deep dives from across the industry to help you choose the right STT API.
A side-by-side look at the leading STT APIs across accuracy, latency, multilingual support, and pricing.
Benchmarks and deployment guidance for Whisper, Canary-Qwen, Parakeet, Moonshine, and other open ASR models.
Independent batch STT rankings comparing accuracy, speed, and price across commercial and open models.
Community benchmark tracking word error rates for open-source automatic speech recognition models.
Overview of top STT models and APIs for developers evaluating accuracy, latency, and language coverage.
Deepgram's ranking of the top commercial STT APIs by accuracy, speed, cost, and customization.
Independent overview of 14 STT providers with guidance on metrics that predict production performance.
Unified comparison of major STT APIs on accuracy, features, pricing, and integration options.
Guide to choosing an STT API for real-time transcription, noise handling, and production voice apps.
Developer-focused roundup of STT APIs suited for AI notetaker and meeting-assistant products.
Benchmark-driven analysis of leading STT APIs for accuracy, latency, and developer experience.
Hands-on review of the best dictation tools and speech-to-text apps for everyday productivity.
Live side-by-side comparison of real-time STT providers including Soniox, OpenAI, Google, and Deepgram.
Curated collection of transcription models on OpenRouter, from Whisper to GPT-4o Transcribe and Chirp 3.