Post Content
Microsoft AI (MAI) just announced a family of new models including new Thinking, Coding, Voice, Transcription, and Image models. In this session, a leader from MAI will share an open and scientific walk through of what it takes to train our models, what we learn along the way, and how those learnings are designed into the model architectures, features, and capabilities.
Seating for this session is first-come, first-served. Add it to your schedule to plan your day and arrive early to secure a spot.
To learn more, please check out these resources:
* https://aka.ms/build26-next-steps
𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀:
* Dave Citron
𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻:
This is one of many sessions from the Microsoft Build 2026 event. View even more sessions on-demand and learn about Microsoft Build at https://build.microsoft.com
DEM323 | English (US) | Working with models
Demo | (200) Intermediate
#MSBuild
Chapters:
0:00 – Session introduction by Dave Citron, CVP of Microsoft AI
00:01:34 – Introduction of Transcribe 1.5, Voice 2, and Code 1 Flash performance highlights
00:05:25 – Launch of Voice 2 speech generation model focusing on natural prosody
00:05:47 – Showcase of fine-grained emotional control and multilingual availability
00:06:24 – Demonstration of Voice 2 emotional tone ‘joy’ playback
00:10:02 – Reference to 100-page technical report detailing model development
00:13:32 – Benchmark performance and real-world code evaluation
00:14:19 – Introduction to Microsoft Frontier Tuning
00:15:59 – Real-world example: Land O’Lakes quality report generation success Read More Microsoft Developer
