← All sources

Microsoft Research news

The latest AI and developer news from Microsoft Research, with short summaries.

Visit website →

MagenticLite, MagenticBrain, Fara1.5: An Agentic Experience Optimized for Small Models

·2026.05.22 02:00

Vega: Zero-Knowledge Proofs for Digital Identity in the AI Era

·2026.05.21 22:00

Further Clarification on AI Delegation and Long-Horizon Task Reliability

·2026.05.16 03:00

mimalloc: A High-Performance Scalable Memory Allocator for Modern Times

·2026.05.14 02:00

GridSFM: A New Small Foundation Model for Power Grids

·2026.05.14 01:00

Advancing Materials AI with MatterSim: Experimental Synthesis, High-Speed Simulation, and a Multi-Task Model

·2026.05.12 22:00

SocialReasoning-Bench: Measuring Whether AI Agents Act Most Favorably for Users

·2026.05.12 02:00

A pipeline for building a large-scale realistic transmission grid dataset from public data

·2026.05.09 04:00

Microsoft at NSDI 2026: Advances in Large-Scale Network Systems

·2026.05.06 01:00

AI Agent Network Red Teaming: Vulnerabilities Revealed in Large-Scale Interactions

·2026.05.01 06:00

AutoAdapt: Automated Domain Adaptation for Large Language Models

·2026.04.23 01:00

Can AI Help Build a More Sustainable World?

·2026.04.21 01:00

Rapid AI-Driven Shifts in Work, but Benefits Are Uneven

·2026.04.10 01:00

Ideas: How to steer AI toward the future of work we want

·2026.04.10 01:00

ADeLe: Predicting and Explaining AI Performance Across Diverse Tasks

·2026.04.02 01:00

AsgardBench, a Benchmark for Visually Grounded Interactive Planning

·2026.03.27 04:00

GroundedPlanBench: Spatially Grounded Long-Horizon Task Planning for Robotic Manipulation

·2026.03.27 01:00

Will Machines Ever Have Intelligence

·2026.03.24 00:00

Systematic Debugging for AI Agents: Introducing the AgentRx Framework

·2026.03.13 01:00

PlugMem: Turning Raw Agent Interactions into Reusable Knowledge

·2026.03.11 01:00