Shengyu Fu

Building AI-assisted developer tools in CoreAI@Microsoft. An engineer, ML&AI scientist and avid runner. @GithubCopilot @code @VisualStudio

[Visual Studio Code @code]

🧠 HydraFusion is now available in @code as a Research Preview.

Instead of relying on a single model, HydraFusion orchestrates multiple models and workflows to tackle coding tasks.

Give it a try today! 🚀

https://t.co/CCB7TN69Rz

[GitHub @github]

👋 Project HydraFusion is now available in the GitHub Copilot app and @code.

Choose it like any other model. Behind the scenes it routes your task across multiple models to draft, critique, revise, or escalate, then hands you one result. https://t.co/CJTClEIVBx

[Andrea @acolombiadev]

HydraFusion in Copilot CLI isn’t another model in the picker. You select it once. It decides whether the task is a single pass, a cheap draft with a quality gate, or draft → critique → revise.
Wrote it up in plain English https://t.co/JSfe0t2lp3

[Satya Nadella @satyanadella]

Super excited about HydraFusion in GitHub Copilot, and what it shows about the shift from model selection to model orchestration.

By bringing together multiple models to plan, build, critique, and complete coding tasks, it can deliver outcomes at up to 67% lower cost.

It’s a great example of the value of a heterogeneous model ecosystem, and how we’re continuing to advance the cost-to-outcome frontier.

[Shengyu Fu @ShengyuFu]

Excited to share HydraFusion, our latest research preview in GitHub Copilot CLI. HydraFusion delivers frontier intelligence through adaptive multi-model orchestration. It dynamically combines models and orchestration patterns to balance quality, cost, and latency.

[GitHub @github]

4.9 percentage points higher verified task quality. 67% lower estimated cost.

Project HydraFusion delivered those results against Claude Opus 5 on Terminal-Bench 2.1 in controlled offline evaluations.

HydraFusion orchestrates the models and workflow for each coding task. You select it like any other model. ⬇️
https://t.co/Vj7R95fXlS

[Pierce Boggan @pierceboggan]

GitHub Copilot gives you access to SOTA models from leading labs.

Auto mode ensures you use the best for each task!

[Kyle Daigle @kdaigle]

GitHub Copilot now has 4.7M+ paid subscribers, up 75% YoY. With Agent HQ and Copilot SDK, we’re building the platform where developers and agents truly work together. 18 years in, the bet on collaboration just keeps growing.

As always... more to come. 😉

[Shengyu Fu @ShengyuFu]

The road to better completions: Building a faster, smarter GitHub Copilot with a new custom model https://t.co/LWDYOUrGFu

[Pierce Boggan @pierceboggan]

our science team wrote up a post detailing how we build the custom models powering completions in @code!

we appreciate your continued feedback to continue this experience better :)

science details: https://t.co/lgMXya9da7

[Mario Rodriguez @mariorod1]

We keep pushing forward making code completion better for all developers. This post details some of the science work behind our model improvements.

https://t.co/XRE2fBGeql

[Mario Rodriguez @mariorod1]

Just today:
1. New embeddings model
2. A new Copilot-SWE model for agentic scenarios AND

[Shengyu Fu @ShengyuFu]

@robertotcestari Not yet, but we'll experiment that soon, stay tuned!

[Shengyu Fu @ShengyuFu]

Had the honor to blog about the new embedding model we just shipped in Github Copilot, love to hear users' feedback!
GitHub Copilot gets smarter at finding your code: Inside our new embedding model  https://t.co/QxxgOJVAab

[Shengyu Fu @ShengyuFu]

@OrenMe @pierceboggan My personal take: Semantic search is one tool in a model’s arsenal. When retrieval quality is high, the agent’s performance improves dramatically, when it’s noisy or stale, it becomes a liability.

[Mario Rodriguez @mariorod1]

As promised ... here is an inside look into our new embeddings model.

https://t.co/WvYL5DOEut

[Pierce Boggan @pierceboggan]

our team has been hard at work to improve the code search experience within GitHub Copilot to give you faster, more accurate code results!

we wrote this blog to share more about our Copilot-Embedding model, and specifically details on the training techniques

if you like posts like this, we'll do more to share our learnings from our science initiatives :)

https://t.co/NLh4ivj9f9

[Shengyu Fu @ShengyuFu]

@tristanbob @mariorod1 @pierceboggan Many retrieval skills transfer well from Python to other languages, so extra Python coverage is an efficient way to boost generalization while we still keep substantial representation for Java, C++, JS/TS, C#, and others.

[Aaron Yim @aaronyim]

We're excited to launch a *public preview* of IntelliCode whole line completions for #Python in @code

🎯 Try it at https://t.co/bAZhzPouKV. We like it already, but it'll keep getting better as we work on it.

Let me know what you think here or at [email protected]

[Shengyu Fu @ShengyuFu]

Shengyu Fu completed the RnR Seattle Marathon! Time 03:43:12 (UNOFFICIAL) https://t.co/4m6jBMv520 #RNRSEATTLE

[Shengyu Fu @ShengyuFu]

Shengyu Fu crossed ANNOUNCER. Time 03:42:55, pace 09:42 min/mile, est. finish 10:10 am https://t.co/4m6jBMv520 #RNRSEATTLE