← Back
VentureBeat AI

Microsoft AI’s MAI-Transcribe-2 undercuts OpenAI, Google and ElevenLabs on price and speed

•12 min read•
#deployment
Microsoft AI’s MAI-Transcribe-2 undercuts OpenAI, Google and ElevenLabs on price and speed
Level:Advanced
For:AI Engineers
✦TL;DR

Microsoft AI unveiled MAI‑Transcribe‑2, a speech‑recognition model priced at 10 ¢ per hour of audio—about a 72 % cut from its prior $0.36‑per‑hour rate—and claims a 5.2 % average word error rate, topping the FLEURS benchmark across 60 languages. The service bundles speaker diarization, word‑level timestamps, keyword biasing, automatic language identification, configurable output styles (verbatim vs. clean), and code‑switching support (e.g., Hinglish, Spanglish). Engineers can access the model via Microsoft Foundry or the MAI Playground, enabling enterprise‑scale transcription at dramatically lower cost.

⚡ Key Takeaways

  • Price reduced to 10 ¢ per hour, a 72 % drop from the previous $0.36 per hour.
  • Supports 60 languages, up from 43 in MAI‑Transcribe‑1.5 and 25 in the original release.
  • Achieves a 5.2 % average word error rate, ranking #1 on the FLEURS benchmark.
  • Includes built‑in speaker diarization, word‑level timestamps, keyword biasing, automatic language ID, configurable output styles, and code‑switching handling.
  • Available through Microsoft Foundry marketplace and MAI Playground testing environment.

🔧 Tools & Libraries

Microsoft FoundryMAI Playground
💡 Why It Matters

The steep price cut and rich feature set make high‑volume, multilingual transcription viable for large enterprises without a separate line‑item cost, accelerating AI‑driven contact‑center and compliance workflows.

✅ Practical Steps

  1. Provision MAI‑Transcribe‑2 from Microsoft Foundry or test it in the MAI Playground.
  2. Enable speaker diarization and word‑level timestamps in the request payload to obtain structured transcripts.
  3. Use the configurable output style flag to select “verbatim” for compliance or “clean” for readable captions.

Want the full story? Read the original article.

Read on VentureBeat AI ↗

More like this

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

AWS ML Blog•#agents

Three insights you may have missed from theCUBE’s coverage of the Dell Technologies New York Symposium

SiliconANGLE AI•#deployment

Welcome Inkling 2 Mini, Cutting-edge Performance for your Local Setup

Hugging Face Blog•#llm

NVIDIA and Microsoft Expand Accelerated Computing From Azure to Windows PCs

NVIDIA Blog•#compute

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING