Microsoft's MAI-Transcribe-2 hits ASR leaderboard at $0.10/hour, far below rivals

ArtificialAnlys · x · 2026-09-03

Artificial Analysis added Microsoft's new MAI-Transcribe-2 speech-to-text model to its non-streaming ASR leaderboard; the model is available via Microsoft Foundry.

Pricing is the standout: $0.10 per hour (about $1.67 per 1,000 minutes), well below comparable high-accuracy rivals — ElevenLabs Scribe v2 ($3.67), Smallest AI Pulse Pro ($4), and Gemini 3.5 Transcribe ($5) per 1,000 minutes.

The AA-WER index averages word error rate across 8 hours of audio from three datasets (AA-AgentTalk 50%, VoxPopuli 25%, Earnings22 25%), covering accents, domain language, and tough acoustic conditions across 67 models.

Related event: Microsoft Launches MAI-Transcribe-2: Faster, Second-Most Accurate, and Cheapest in Class(7 posts)→

Original post →

More from Models

Models channel →