Microsoft Research's banner

Microsoft Research

@MSFTResearch • 554,764 subscribers

We advance science and technology to benefit humanity.

Shorts

Microsoft researchers introduce MatterGen, a model that can discover new materials tailored to specific needs—like efficient solar cells or CO2 recycling—advancing progress beyond trial-and-error experiments.

Microsoft researchers introduce MatterGen, a model that can discover new materials tailored to specific needs—like efficient solar cells or CO2 recycling—advancing progress beyond trial-and-error experiments.

1,329,435 次观看

LLMs do not get smarter just by remembering more. EvoLib turns experience into evolving knowledge, taking reusable skills and insights that help models learn and adapt across tasks long after deployment.

LLMs do not get smarter just by remembering more. EvoLib turns experience into evolving knowledge, taking reusable skills and insights that help models learn and adapt across tasks long after deployment.

23,059 次观看

Cryptographic code supports vital protections in modern computing systems. Learn how a new method helps verify code as developers write it while preserving speed and adaptability as it gets implemented and evolves.

Cryptographic code supports vital protections in modern computing systems. Learn how a new method helps verify code as developers write it while preserving speed and adaptability as it gets implemented and evolves.

23,154 次观看

AI agents often fail because their instructions, or skills, are manually modified with no guarantee of improvement. Learn how SkillOpt turns skill editing into a training process, making agent behavior more reliable without changing model weights:

AI agents often fail because their instructions, or skills, are manually modified with no guarantee of improvement. Learn how SkillOpt turns skill editing into a training process, making agent behavior more reliable without changing model weights:

20,451 次观看

Aurora 1.5 adds 22 more variables, hourly temporal resolution, and probabilistic ensemble forecasting to the Aurora foundation model, making it more useful for real-world weather, climate, and energy applications.

Aurora 1.5 adds 22 more variables, hourly temporal resolution, and probabilistic ensemble forecasting to the Aurora foundation model, making it more useful for real-world weather, climate, and energy applications.

13,145 次观看

Vega turns a full credential into a single proof, sharing only what is needed and nothing more, with performance that works in real apps.

Vega turns a full credential into a single proof, sharing only what is needed and nothing more, with performance that works in real apps.

18,083 次观看

Safe agents don’t guarantee a safe ecosystem of interconnected agents. Microsoft Research examines what breaks when AI agents interact and why network-level risks require new approaches. Learn more:

Safe agents don’t guarantee a safe ecosystem of interconnected agents. Microsoft Research examines what breaks when AI agents interact and why network-level risks require new approaches. Learn more:

20,174 次观看

Introducing GridSFM, a small foundation model that can predict AC optimal power flow in milliseconds, boosting efficiency and unlocking cost savings. Learn how GridSFM gives grid operators direct visibility into congestion, stability, and system health:

Introducing GridSFM, a small foundation model that can predict AC optimal power flow in milliseconds, boosting efficiency and unlocking cost savings. Learn how GridSFM gives grid operators direct visibility into congestion, stability, and system health:

15,206 次观看

MatterSim is expanding what AI can do for materials science—from faster large-scale simulations to MatterSim-MT, a new multi-task model for simulating properties beyond potential energy surfaces alone.

MatterSim is expanding what AI can do for materials science—from faster large-scale simulations to MatterSim-MT, a new multi-task model for simulating properties beyond potential energy surfaces alone.

14,707 次观看

Using SocialReasoning Bench, we observed a stable pattern across models—agents execute competently, but fail to consistently improve the user’s position, even with explicit instructions to optimize for user interest.

Using SocialReasoning Bench, we observed a stable pattern across models—agents execute competently, but fail to consistently improve the user’s position, even with explicit instructions to optimize for user interest.

12,442 次观看

AutoAdapt automates the design and tuning of domain adaptation workflows for large language models. It improves performance without requiring additional compute, making deployment more accessible:

AutoAdapt automates the design and tuning of domain adaptation workflows for large language models. It improves performance without requiring additional compute, making deployment more accessible:

10,753 次观看

In Episode 1 of the podcast “The AI Revolution in Medicine, Revisited,” host Peter Lee, president of Microsoft Research, & guests Dr. Christopher A. Longhurst, MD & Dr. Sara Murray discuss how generative AI is impacting patient care & workflows in healthcare:

In Episode 1 of the podcast “The AI Revolution in Medicine, Revisited,” host Peter Lee, president of Microsoft Research, & guests Dr. Christopher A. Longhurst, MD & Dr. Sara Murray discuss how generative AI is impacting patient care & workflows in healthcare:

18,482 次观看

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Coming May 14 at Microsoft Research Forum: a new release and demo from MSR AI Frontiers. Plus new work on Agentic GitHub Workflows, Real-time agent verification, Energy-based fine-tuning, and Guiding the AI transition. Register now:

Coming May 14 at Microsoft Research Forum: a new release and demo from MSR AI Frontiers. Plus new work on Agentic GitHub Workflows, Real-time agent verification, Energy-based fine-tuning, and Guiding the AI transition. Register now:

Microsoft Research

17,748,463 次观看 • 4 个月前

The latest Microsoft Research Forum episode is now available on-demand. Explore new ARO, Dion2, Magentic Marketplace, OptiMind, Agent Lightning, and Healthbots. Register to watch:

The latest Microsoft Research Forum episode is now available on-demand. Explore new ARO, Dion2, Magentic Marketplace, OptiMind, Agent Lightning, and Healthbots. Register to watch:

Microsoft Research

21,197,502 次观看 • 6 个月前

New tools, models, repos, and papers out of Microsoft Research are here. Use AI and agents? It's worth watching: • MagenticLite from MSR AI Frontiers • Agentic GitHub workflows • Verification‑first agents • Meaning‑matching fine‑tuning • AI transition + the economy

New tools, models, repos, and papers out of Microsoft Research are here. Use AI and agents? It's worth watching: • MagenticLite from MSR AI Frontiers • Agentic GitHub workflows • Verification‑first agents • Meaning‑matching fine‑tuning • AI transition + the economy

Microsoft Research

3,447,138 次观看 • 3 个月前

AI is transforming the way we all live, work, and think. Join us for Microsoft Research Forum, where you'll learn about the most recent advances and hear bold new ideas from the global research community. See you on June 4 for Episode 3. Register now.

AI is transforming the way we all live, work, and think. Join us for Microsoft Research Forum, where you'll learn about the most recent advances and hear bold new ideas from the global research community. See you on June 4 for Episode 3. Register now.

Microsoft Research

22,477,328 次观看 • 2 年前

Today in the journal Science: BioEmu from Microsoft Research AI for Science. This generative deep learning method emulates protein equilibrium ensembles – key for understanding protein function at scale.

Today in the journal Science: BioEmu from Microsoft Research AI for Science. This generative deep learning method emulates protein equilibrium ensembles – key for understanding protein function at scale.

Microsoft Research

415,410 次观看 • 1 年前

Orchard is an open-source framework for the research community to train and evaluate AI agents across task types. It reduces complexity while supporting strong performance from smaller models by enabling researchers to reuse the same infrastructure.

Orchard is an open-source framework for the research community to train and evaluate AI agents across task types. It reduces complexity while supporting strong performance from smaller models by enabling researchers to reuse the same infrastructure.

Microsoft Research

25,995 次观看 • 1 个月前

Radiology AI is evolving beyond report generation. CARE-X explores a unified approach that combines flexible reasoning, calibrated predictions, and measurement-based tools for chest X-ray interpretation.

Radiology AI is evolving beyond report generation. CARE-X explores a unified approach that combines flexible reasoning, calibrated predictions, and measurement-based tools for chest X-ray interpretation.

Microsoft Research

17,558 次观看 • 24 天前

LLMs do not get smarter just by remembering more. EvoLib turns experience into evolving knowledge, taking reusable skills and insights that help models learn and adapt across tasks long after deployment.

LLMs do not get smarter just by remembering more. EvoLib turns experience into evolving knowledge, taking reusable skills and insights that help models learn and adapt across tasks long after deployment.

Microsoft Research

23,059 次观看 • 1 个月前

Every powerful new technology comes with trade-offs. The internet had them. Social media had them. Doug Burger, Amy Luers, and Ishai Menache explore AI’s climate impact and steering the tech to support a sustainable future.

Every powerful new technology comes with trade-offs. The internet had them. Social media had them. Doug Burger, Amy Luers, and Ishai Menache explore AI’s climate impact and steering the tech to support a sustainable future.

Microsoft Research

11,599 次观看 • 17 天前

Small language models learn to negotiate with SocialRL, PazaBench V2 expands speech AI evaluation across African languages, and EvoLib helps agents turn experience into knowledge. Plus: new methods for more reliable A/B testing and advances in AI-driven precision oncology.

Small language models learn to negotiate with SocialRL, PazaBench V2 expands speech AI evaluation across African languages, and EvoLib helps agents turn experience into knowledge. Plus: new methods for more reliable A/B testing and advances in AI-driven precision oncology.

Microsoft Research

19,104 次观看 • 1 个月前

The wait is over! Microsoft Research is sharing Skala, the new exchange-correlation functional, marking a major milestone in the accuracy/cost trade-off in DFT. Help us learn from your testing so we can improve. Available on Azure AI Foundry and GitHub.

The wait is over! Microsoft Research is sharing Skala, the new exchange-correlation functional, marking a major milestone in the accuracy/cost trade-off in DFT. Help us learn from your testing so we can improve. Available on Azure AI Foundry and GitHub.

Microsoft Research

127,370 次观看 • 11 个月前

Microsoft Chief Scientific Officer Eric Horvitz on what stood out at NeurIPS 2025, what’s next, and why these moments matter for the future of AI research.

Microsoft Chief Scientific Officer Eric Horvitz on what stood out at NeurIPS 2025, what’s next, and why these moments matter for the future of AI research.

Microsoft Research

86,914 次观看 • 8 个月前

Cryptographic code supports vital protections in modern computing systems. Learn how a new method helps verify code as developers write it while preserving speed and adaptability as it gets implemented and evolves.

Cryptographic code supports vital protections in modern computing systems. Learn how a new method helps verify code as developers write it while preserving speed and adaptability as it gets implemented and evolves.

Microsoft Research

23,154 次观看 • 1 个月前

MagenticLite's models are now fully open source. MagenticBrain and Fara 1.5, previously available on Microsoft Foundry, are live on Hugging Face with open weights. The app, the harness, and every model in the stack are now open.

MagenticLite's models are now fully open source. MagenticBrain and Fara 1.5, previously available on Microsoft Foundry, are live on Hugging Face with open weights. The app, the harness, and every model in the stack are now open.

Microsoft Research

19,294 次观看 • 1 个月前

Today, we welcome the 2026 Microsoft Research Fellowship cohort, an inspiring global community of fellows and advisors helping to shape what’s next across science, technology, and society. Join us in celebrating this year’s recipients: These contributions span the following themes: • AI for global and societal impact • AI fundamentals: scalable reasoning, model adaption and evaluation • Biological and scientific modeling • Foundational systems & infrastructure for AI • Human-AI collaboration and interaction • Multimodal & embodied intelligence

Today, we welcome the 2026 Microsoft Research Fellowship cohort, an inspiring global community of fellows and advisors helping to shape what’s next across science, technology, and society. Join us in celebrating this year’s recipients: These contributions span the following themes: • AI for global and societal impact • AI fundamentals: scalable reasoning, model adaption and evaluation • Biological and scientific modeling • Foundational systems & infrastructure for AI • Human-AI collaboration and interaction • Multimodal & embodied intelligence

Microsoft Research

49,992 次观看 • 4 个月前

Aurora 1.5 scales open forecasting; Flint redefines visualization; FlowDAgger accelerates adaptation. Plus: AI agents crack Rowhammer defenses and memory reshapes conversational AI.

Aurora 1.5 scales open forecasting; Flint redefines visualization; FlowDAgger accelerates adaptation. Plus: AI agents crack Rowhammer defenses and memory reshapes conversational AI.

Microsoft Research

16,072 次观看 • 1 个月前

AI agents often fail because their instructions, or skills, are manually modified with no guarantee of improvement. Learn how SkillOpt turns skill editing into a training process, making agent behavior more reliable without changing model weights:

AI agents often fail because their instructions, or skills, are manually modified with no guarantee of improvement. Learn how SkillOpt turns skill editing into a training process, making agent behavior more reliable without changing model weights:

Microsoft Research

20,451 次观看 • 2 个月前

Researchers Jenna Butler, Jake Hofman, and Rebecca Janssen spend a lot of time studying AI’s influence on how people do their jobs. On the Microsoft Research Podcast, they share their thoughts on what the future of work looks like in the age of AI.

Researchers Jenna Butler, Jake Hofman, and Rebecca Janssen spend a lot of time studying AI’s influence on how people do their jobs. On the Microsoft Research Podcast, they share their thoughts on what the future of work looks like in the age of AI.

Microsoft Research

11,071 次观看 • 1 个月前

Aurora 1.5 adds 22 more variables, hourly temporal resolution, and probabilistic ensemble forecasting to the Aurora foundation model, making it more useful for real-world weather, climate, and energy applications.

Aurora 1.5 adds 22 more variables, hourly temporal resolution, and probabilistic ensemble forecasting to the Aurora foundation model, making it more useful for real-world weather, climate, and energy applications.

Microsoft Research

13,145 次观看 • 1 个月前