Microsoft Launches Seven New MAI Models
Microsoft launches new self-developed MAI models, including the MAI-Thinking-1 reasoning model, MAI-Image-2.5 image model, MAI-Image-2.5-Flash efficient image model, MAI-Voice-2 voice model, MAI-Voice-2-Flash efficient voice model, MAI-Transcribe-1.5 speech-to-text model, and MAI-Code-1-Flash coding model. All models are trained from scratch without distillation and share unified data specifications and evaluation frameworks.
Microsoft launches new self-developed MAI models, including the MAI-Thinking-1 reasoning model, MAI-Image-2.5 image model, MAI-Image-2.5-Flash efficient image model, MAI-Voice-2 voice model, MAI-Voice-2-Flash efficient voice model, MAI-Transcribe-1.5 speech-to-text model, and MAI-Code-1-Flash coding model. All models are trained from scratch without distillation and share unified data specifications and evaluation frameworks.
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.