On Thursday, Microsoft released the MAI-Transcribe-2 speech-to-text model. This model outperforms existing products from OpenAI, Google, and ElevenLabs in terms of speed and accuracy, and it is available at a price of $0.1 per hour. Compared to the prices of similar models five months ago (which were $0.36 per hour), this early-bird pricing reduction reduces costs by approximately 72%. For businesses that process 100,000 calls per year, the annual cost drops from $36,000 to $10,000. This release reflects Microsoft’s strategy of building cutting-edge models in the field of speech transcription and gradually integrating them into its products.
On Thursday, Microsoft AI released MAI-Transcribe-2, a voice-to-text model that is faster, more accurate, and cheaper than the products currently available from OpenAI, Google, and ElevenLabs. The model is sold for $0.1 per hour. When Microsoft AI launched this series of models five months ago, the price was $0.36 per hour. The early bird pricing reduction on Thursday reduced the price by approximately 72%. For businesses that process 100,000 calls per year, the annual cost drops from $36,000 to $10,000. At this price level, transcription is no longer a topic of discussion for businesses. This release reflects Microsoft’s strategy of building cutting-edge models in one domain and gradually integrating them into its products, with transcription being the area where this strategy is advancing fastest. MAI-Transcribe-2 also demonstrates how Microsoft can develop advanced models without relying on its partners, who have invested $13 billion in development…