#microsoft

News · 2026-11-05

PDD Kit moves review of agent-written code from lines to specifications

Microsoft Research and GitHub Next demonstrated a prototype that turns developer intent into a Verus specification and builds machine-checked proofs for Rust code. The stronger assurance applies only inside an explicitly defined trust boundary.

News · 2026-11-03

Cybersecurity runs on trust, not blind rule enforcement

Microsoft Research's Mike Walker argues that every security demand spends from a limited reserve of user trust. His experience with DARPA's Cyber Grand Challenge also shows why AI defense needs team coordination and context rather than another solitary supertool.

News · 2026-10-30

Microsoft tests Copilot on synthetic companies instead of customer data

Microsoft Research built three synthetic enterprise environments for testing Copilot without using customer data. Its internal quality metric improved by 39% over the initial approach, but a realistic simulation is still not proof of performance inside a live company.

News · 2026-10-29

Microsoft Research Focus 97 Maps Agent Bottlenecks, Forecast Bias and Real Workplace Use

In Edition 97 of Microsoft Research Focus, published on October 29, 2026, Microsoft brings together multiple projects spanning agent efficiency, coding benchmarks, workplace telemetry, Indian classrooms, subseasonal weather forecasting, and the BioEmu2 public preview. Read as a research map, the roundup shows that practical AI progress increasingly depends on measuring specific failure modes, from unfamiliar codebases and dominant office habits to systematic forecast bias.

News · 2026-10-22

Microsoft wants AI agents to draw charts you can actually click

Microsoft Research has extended its Flint chart language with interactions: the agent describes what should happen on a click or selection, and the compiler handles the rest. A selection in the chart can also flow back into the chat as context for the next question. The claims about lower cost and higher reliability come without numbers so far.

News · 2026-10-20

The man behind DARPA's Memex now runs Microsoft's Catalyst Lab, and he is betting on cyber defense

Microsoft Research has released a conversation with Chris White, head of the new Catalyst Lab. At DARPA, White led Memex, dark web search tools used against human trafficking. He now wants to apply the same dot-connecting logic to network defense, but has not yet shown a concrete project.

News · 2026-10-21

Mathematicians with AI agents proved the 31-year-old KLS conjecture. Unguided models with the same tokens did not

Xinzhi Zhang of Microsoft Research and co-author Zhao Song have proved the Kannan, Lovász and Simonovits conjecture from 1995. AI agents explored more than 100 candidate approaches, but humans set the direction, and a control attempt without them fell short on a similar token budget.

News · 2026-10-15

Magnetar speeds up proofs for ordinary code without erasing the trust bill

Microsoft Research has presented Magnetar, a zkVM whose roughly 600 KiB proofs verified in 10 to 12 ms on an M4 Max. It brings verifiable computation closer to ordinary programs, but the evidence still rests mainly on the authors’ benchmarks.

News · 2026-10-12

StudentSim tests AI tutors on digital replicas of learners

Microsoft Research introduced StudentSim, a framework that trains personalized learner simulators across chess, ESL writing, and mathematics. The system provides immediate feedback for tutoring models without testing unverified hints on real students.

News · 2026-10-07

Agent Lightning trains agents in their real harness with 3,500 lines of code

Microsoft has introduced Agent Lightning v1.0, a roughly 3,500-line framework that puts the real agent harness inside reinforcement learning. With 6,000 training examples, it raised Qwen3.5-9B on SWE-bench Verified from 41.8% to 56.4%.

News · 2026-10-06

A benchmark praises the model, a conversation lets it get lost

Microsoft Research's Jennifer Neville argues that common AI benchmarks compress work into one perfectly specified instruction. In real conversations, people reveal requirements gradually, and her research finds that model performance drops substantially.

News · 2026-09-30

A model forecasts risk for 66,935 substations up to an hour before a solar storm hits

Microsoft Research built a system that estimates space-weather risk 30 to 60 minutes ahead for 66,935 substations across the continental United States. It detected 76.5% of major events in testing, but still needs validation with utilities and operational data before deployment.

News · 2026-09-29

Quine connects biological data to narrow the experimental queue

Microsoft Research introduced Quine, an early research system combining a multimodal biology world model with tools for choosing experiments. An initial project with the Broad Institute ranked thousands of compounds for shifting cell states in pancreatic cancer.

News · 2026-09-28

Microsoft is building a bridge from AI papers to regulated industry in Singapore

The first year of Microsoft Research Asia's Singapore lab produced robot vision research and partnerships with universities, healthcare, and government. Its real regional metric will be systems validated beyond the conference hall.

News · 2026-09-23

Remote GPUs help robots, but the network controls their reflexes

A Microsoft study found that moving inference from a robot to edge or cloud GPUs improves performance and battery life. It also exposes the hard constraint: tens of milliseconds of network delay already reduce the accuracy of physical work.

News · 2026-09-21

Chimera framework advances drug discovery by ensembling diverse AI models

Microsoft Research and Novartis have introduced Chimera, a learning-to-rank system that shortens the retrosynthesis process in drug development.

News · 2026-07-30

EvoLib replaces memory accumulation in LLMs with inference-time skill extraction

Microsoft Research introduced EvoLib, a framework that distills general skills from raw conversation history at inference time. This allows agents to improve runtime performance without requiring model retraining.

News · 2026-07-30

Microsoft Echoverse builds evolving environments for agents, not static tests

Current agents fail in multi-step processes. Instead of static benchmarks, Microsoft Research trains them in simulations that dynamically evolve alongside task complexity.

News · 2026-08-31

Microsoft’s GigaPath-Flash cuts the compute cost of pathology foundation models by 50x

Microsoft has distilled its massive pathology models into the Flash family. This delivers a 50x reduction in compute cost for whole-slide analysis, unlocking the possibility of repeated experiments across large patient populations.

News · 2026-08-20

Skala 1.1 opens deep learning for quantum chemistry

Microsoft has released an update to its Skala model for quantum chemistry calculations. Version 1.1 makes more accurate molecular simulations accessible to a broader developer ecosystem and adds a living benchmark.

News · 2026-08-12

Microsoft tests spatial reasoning: The new MindTopo benchmark

Microsoft introduced MindTopo, a benchmark for testing the spatial and topological reasoning of visual models. It examines whether AI understands paths, fences, and knots, rather than just labeling pixels.

News · 2026-08-11

Microsoft adds precise measurement tools to Qwen for radiology

The Qwen3-VL-4B-Instruct model in Microsoft's research pipeline no longer just visually inspects X-rays, but triggers external tools to calculate dimensions. For threshold-dependent diagnostics, this means the end of merely guessing shapes.

News · 2026-08-03

Microsoft opens Orchard: a greenhouse for training agents, not another orchestrator

Microsoft Research released Orchard, an open-source framework for scalable agentic model training. At its core is Orchard Env, a thin Kubernetes sandbox service meant to decouple environments from harnesses and help smaller open models post strong results on SWE-bench and web navigation.

News · 2026-07-13

Microsoft moves Lean-checked crypto proofs into SymCrypt production code

Microsoft has published a SymCrypt branch with formal specifications and proofs for Rust implementations of SHA-3 and ML-KEM. The important move is not AI writing proofs, but proofs living next to production cryptographic code.

News · 2026-06-29

Memora tackles agent memory by separating storage from retrieval

Microsoft Research describes Memora as a memory system for AI agents that currently need to reload or retrieve context repeatedly. The important move is separating what gets stored from how it is later retrieved.

News · 2026-07-09

Aurora 1.5 pushes AI weather models toward energy and climate operations

Microsoft Research says Aurora 1.5 adds 22 more variables, hourly resolution and probabilistic ensemble forecasting. The interesting part is not prettier weather maps, but whether AI forecasts can support expensive operational decisions.

News · 2026-07-08

Flint gives AI agents a shorter path from data to charts

Microsoft Research is showing Flint, an open-source visualization language that lets AI agents build charts from compact, human-editable specs. The interesting part is not another charting framework, but the layer between a prompt and the rendering library.

News · 2026-07-02

Microsoft's Ire found LOTUSLITE where signatures failed

Microsoft showed Project Ire classifying a 253 KB LOTUSLITE DLL variant as malware even though VirusTotal showed only 1 of 72 vendors detecting it on May 28, 2026. The important part is not attribution, but an agent reading behavior instead of matching IOC lists.

News · 2026-07-02

Microsoft turns AI explainability into a hypothesis engine for neuroscience

Microsoft Research described generative causal testing, a method that turns brain prediction models into readable hypotheses and checks them with fMRI. The interesting move is not prediction alone, but using AI to return testable theory to scientists.

News · 2026-06-30

SkillOpt trains agent skills like text weights

Microsoft is showing SkillOpt, an optimizer that improves an agent skill file without changing model weights. For teams building agents, the important part is the validation gate, not another layer of prompt mysticism.