Microsoft Research published a paper showing that a 4-billion-parameter AI model trained with SocialRL, a cascade reinforcement learning approach, achieved an average utility score of 0.627 across six…
Microsoft Research's Webwright lets AI web agents write command-line programs instead of clicking through tasks, boosting the same GPT-5.4 model's success rate on long-horizon tasks from 33.5% to 60.1…
Recordari, a custom memory graph built by developer Corby, scored 0.831 on the LoCoMo benchmark, outperforming Microsoft Research's Memora, which published a score of 0.801. The benchmark, run with th…
Microsoft Research introduced CARE-X, a unified chest X-ray vision-language model designed to improve clinical usefulness by combining generative report generation with calibrated diagnostic predictio…
Microsoft has open-sourced BitNet, an inference framework for 1-bit LLMs that can run a 100-billion-parameter model on a single CPU at speeds comparable to human reading (5-7 tokens per second). The f…
C1 Bridge, a product from C1 that extends AI Access Management to MCP servers in private networks, uses long-lived TLS connections that require periodic re-authentication to allow customers to revoke …
A Heimwall security team analysis of 27,075 real developer AI prompts from the Microsoft Research MSR 2024 Mining Challenge dataset found two valid OpenAI API keys, one Google API key, and 49 personal…
Perforated AI released a perforated version of ResNet-18 on Hugging Face that matches ResNet-34's performance while maintaining ResNet-18's smaller size and compute efficiency, using dendritic optimiz…
GitHub launched Agentic Workflows into public preview on June 11, 2026, enabling repositories to run coding agents inside GitHub Actions for automated tasks like issue triage. The feature, built by Gi…
Microsoft Research on August 4th spotlighted PRISM2, a multimodal pathology model developed by Paige and Microsoft researchers, trained on 2,350,518 whole-slide images from 685,507 specimens and about…
Researchers from Microsoft Research and Paige, now part of Tempus, developed PRISM2, a pathology foundation model trained on tissue images and language from real pathology reports, and reported in Nat…
Microsoft Research released Orchard, an MIT-licensed framework for training and evaluating AI agents, on August 3rd, aiming to reduce the high costs of building agent infrastructure. The framework inc…
Microsoft Research released Orchard, an open-source framework for training and evaluating AI agents, on August 3, featuring Orchard Env, a Kubernetes-native environment service. The framework reports …
A UC Berkeley study found that experienced open-source developers using AI coding assistants were 19% slower than those working without them, as the overhead of reviewing and correcting AI-generated c…
Microsoft Research released Flint, an open-source intermediate visualization language that compiles AI agents' chart intents into Vega-Lite, Apache ECharts, Chart.js, Plotly, or Excel charts, with ver…
Microsoft Research released Flint, a chart specification language for AI agents that uses a deterministic compiler instead of a larger LLM to generate polished visualizations. The language reduces cha…
Microsoft Research released EvoLib on July 30th, a test-time learning framework that enables large language model agents to improve across tasks without updating their underlying weights. The framewor…
A 2025 Microsoft Research study of 319 knowledge workers found that greater confidence in generative AI was associated with less critical-thinking effort, and a 2026 study in Scientific Reports showed…
Microsoft Research has introduced SkillOpt, a system that enables AI agents to self-evolve by optimizing their skill documents—markdown files that guide agent behavior—without modifying model weights.…
A developer known as slvDev has run a 28.9 million parameter language model on an ESP32-S3 microcontroller that costs about $8, generating text at roughly 9.5 tokens per second entirely on-device. The…