Microsoft AI Updates: April 9, 2026

1. Foundry Labs April Roundup: Phi-4-Reasoning-Vision, MAI Speech Models, and VibeVoice

Microsoft. Microsoft published its Foundry Labs April roundup featuring several models now in public preview: Phi-4-Reasoning-Vision-15B for visual reasoning (88.2% on ScreenSpot_v2), MAI-Transcribe-1 for speech recognition across 25 languages at 3.9% WER with 50% lower GPU cost, MAI-Voice-1 for text-to-speech generating 60 seconds of audio in under one second on a single GPU, MAI-Image-2 (ranked #3 on Arena.ai leaderboard), and VibeVoice ASR for 60-minute single-pass transcription across 50+ languages. The breadth reflects Microsoft’s push to fill model gaps across modalities on the Foundry platform. Source

2. GigaTIME: Multimodal AI for Population-Scale Tumor Microenvironment Analysis

Microsoft. Microsoft Research, in collaboration with Providence and the University of Washington, released GigaTIME on Foundry, a multimodal AI model that translates routine H&E pathology slides into virtual multiplex immunofluorescence images across 21 protein channels. Trained on 40 million cells, GigaTIME was applied to 14,256 cancer patients across 51 hospitals and identified 1,234 statistically significant associations, enabling population-scale tumor immune microenvironment analysis without expensive lab techniques. Source