GitHub
GitHub is the venue for Microsoft's MAI Code 2 and MAI Code 2 Flash models in Copilot, and a partner with Microsoft on an AI vulnerability pilot that produced 99 verified reports from 9,791 detected issues.
Two stories from November 2026 put GitHub at the center of Microsoft's AI tooling for developers. On November 17, Microsoft released the MAI Code 2 and MAI Code 2 Flash programming models, and both are available in GitHub Copilot. Microsoft reports MAI Code 2 at a 63.2% average benchmark success rate, against 64.1% for OpenAI's GPT-6.1 Sol and 65.3% for Anthropic's Claude Opus 5.5, at about 40% of GPT-6.1 Sol's per-task cost. The pricing gap is the point: the model lands within a point of the leaders on Microsoft's own benchmarks while costing far less per task.
The same day, GitHub and Microsoft announced the results of a pilot that uses AI to find vulnerabilities in open source software, then delivers only verified findings and proposed fixes to maintainers. Six projects were examined and 9,791 issues were detected, narrowed to 99 reports. As of November 15, 55 had been fixed and 14 of those fixes were already in releases. The ratio between detected issues and verified reports is the load-bearing number: the bottleneck has moved from finding bugs to verifying and fixing them, and GitHub's own admission that the verification step does not scale yet matters more than the 9,791.
Read together, the two stories describe a company shipping cheaper coding models into Copilot while running an AI-assisted security pipeline whose verification stage is not yet keeping pace with detection. Both announcements are dated November 17, 2026.
Key facts
- Models available in GitHub Copilot
- MAI Code 2 and MAI Code 2 Flash ↗
- MAI Code 2 benchmark success rate
- 63.2% average ↗
- GPT-6.1 Sol benchmark success rate
- 64.1% ↗
- Claude Opus 5.5 benchmark success rate
- 65.3% ↗
- MAI Code 2 per-task cost relative to GPT-6.1 Sol
- About 40% ↗
- AI vulnerability pilot projects examined
- Six ↗
- Issues detected in the pilot
- 9,791 ↗
- Verified reports delivered to maintainers
- 99 ↗
- Fixes in place as of November 15
- 55 fixed, 14 already in releases ↗
Timeline
- Nov 18, 2026 GitHub and Microsoft Cut 9,791 AI-Found Flaws to 99 Verified Reports
- Nov 18, 2026 Microsoft Releases MAI Code 2 and MAI Code 2 Flash for GitHub Copilot
Written with AI from the cited Yomimono stories below, each itself sourced. See our editorial policy. Every fact links the story it came from.
Facts
- Announced
- results of a pilot that uses AI to find vulnerabilities in open source software, then delivers only verified findings and proposed fixes to maintainers (November 17) · 2026-11-18
Structured graph also available as JSON at /public/entities/github.
CC BY 4.0.