Shengyu Fu (@ShengyuFu) header
Shengyu Fu (@ShengyuFu) profile picture

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

Redmond, WA
Joined January 2012
Shengyu Fu retweeted
You shouldn't have to pick a model for every task. Copilot Fusion does it for you. As @SatyaNadella just shared at #MSIgnite, Copilot Fusion now ranks as the #1 coding agent on Terminal-Bench, scoring 5 points ahead of Claude Code at less than half the cost.
🆕 On Terminal-Bench 4.1, Copilot Fusion Max achieved a 71.2% resolution rate, 5.4 percentage points ahead of leading frontier models Opus 5.5 and Fable 5.2, while costing 88% less than Fable 5.2 and 57% less than Opus 5.5. docs.github.com/copilot/conc…
3
3
1
19
1,244
Shengyu Fu retweeted
Instead of picking the model for each and every task, let Copilot Fusion route and orchestrate which model to use. ✅ Copilot Fusion balances performance and cost efficiency, picking from three execution patterns for the job: Single, Cascade, or Critique. All you need to do is choose the effort level. Try it in the GitHub Copilot app, Copilot CLI, and @code, now 30% off through November 30.
58
38
14
356
52,773
Shengyu Fu retweeted
The advantage of running across multiple model families is structural. Agentic iteration converges faster and is more successful. The question was if you can make prompt caching to work well enough - here is the answer:
GitHub has released Copilot Fusion in GitHub Copilot, which orchestrates between Anthropic and OpenAI models. It scores 2 points below Opus 5.5 in Claude Code at half the Cost per Task in the Coding Agent Index GitHub describes Copilot Fusion as a runtime orchestration system, which selects and coordinates models to optimize performance, cost, and latency for each task. Copilot Fusion chooses between a number of models from Anthropic and OpenAI, including Claude Opus 5.5, Claude Sonnet 5.5, GPT-6 Astra, and GPT-6 Sol. We benchmarked its max setting in GitHub Copilot CLI in the Artificial Analysis Coding Agent Index v2.0. Token costs are charged based on the underlying models used to complete a request. Key takeaways: ➤ Cost-efficient for near-frontier performance: Copilot Fusion in GitHub Copilot costs $5.59 per task, versus $11.42 for Claude Code with Opus 5.5 (max). Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high) scores 62 at $8.00 per task. ➤ Slight lead in Terminal-Bench 4.0: The Coding Agent Index v2.0 weights its three constituent benchmarks equally. Copilot Fusion in GitHub Copilot leads Terminal-Bench 4.0 with a score of 71. It ranks third on FrontierCode in the chart at 51, behind Claude Code with Opus 5.5 (max) at 56 and Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high) at 54. Its SWE-Atlas-QnA-AA score lags at 57, between GPT-6 Astra in Codex and Gemini 4 Argon in Antigravity CLI. ➤ Half the token use of Opus 5.5 in Claude Code: GitHub Copilot Fusion used 7.2M tokens per task in the Coding Agent Index, approximately half of the tokens used by Claude Opus 5.5 in Claude Code, and 30% less than Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high).
4
4
18
1,588
1/ We announced our biggest update to Copilot to date, bringing together Chat, Cowork, Autopilot, and Code into one unified experience: · Daily users grew nearly 50% QoQ, and more than 5X YoY. · User satisfaction improved nearly 20% QoQ to another all-time high · Number of customers with more than 50K Copilot seats, up 8X YoY Great momentum with GitHub Copilot too, including Copilot Fusion, announced just this morning, which delivers frontier level quality, at up to 77% lower cost. And this is just the beginning. The next frontier is maximizing intelligence per task while continuously optimizing for cost and performance. This is the frontier that will matter to companies! And we are extending our vision of Copilot as the new OS for work, bringing the business logic and capabilities of D365 directly into Copilot, helping every organization stand up their own infinite SaaS factory.
23
60
14
566
127,521
Shengyu Fu retweeted
Frontier performance at a fraction of the cost. In the @ArtificialAnlys Coding Agent Index, Copilot Fusion Max scored within two points of Claude Code with Opus 5.5 Max at about half the API cost per task. Get 30% off through Nov. 30. Try it now. ⬇️ docs.github.com/copilot/conc…
145
76
96
668
133,729
Shengyu Fu retweeted
🧠 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! 🚀 aka.ms/VSCode/Hydrafusion
23
85
6
662
52,934
Shengyu Fu retweeted
👋 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. github.blog/changelog/2026-0…
72
97
38
539
112,342
Shengyu Fu retweeted
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
3
17
2
103
21,988
Shengyu Fu retweeted
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.
136
192
35
1,431
330,582
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.
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. ⬇️ github.blog/ai-and-ml/github…
1
1
2
438
GitHub Copilot gives you access to SOTA models from leading labs. Auto mode ensures you use the best for each task!
I was skeptical of @github Copilot auto mode...Then Nhu explained how the routing actually works and I can't go back. Turns out a system that picks the right model every time doesn't just save cost, it outperforms using one model alone. give auto another shot 👇🏽youtu.be/a-c1M8QwK84?si=gYjF…
7
5
56
8,902
Shengyu Fu retweeted
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. 😉
12
12
2
214
16,135
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: github.blog/ai-and-ml/github…
4
12
66
4,867
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  github.blog/news-insights/pr…
4
1
13
1,075
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 :) github.blog/news-insights/pr…
8
16
197
23,201
Shengyu Fu retweeted
We're excited to launch a *public preview* of IntelliCode whole line completions for #Python in @code 🎯 Try it at aka.ms/vscode-intellicode-co…. 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]
3
7
2
30