Microsoft's MAI Code 2 claims 96% of Opus 5.5 at 5% of the cost
Two new Microsoft coding models land in GitHub Copilot, benchmarked against GPT-6.1 Sol and Claude Opus 5.5 on cost rather than raw score.
Microsoft released two coding models, MAI Code 2 and MAI Code 2 Flash, on 17 November, and both are now selectable inside GitHub Copilot.
On its own benchmark suite (DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1), Microsoft says MAI Code 2 reaches 98% of GPT-6.1 Sol’s performance at 40% of the cost, and 96% of Claude Opus 5.5’s performance at 5% of the cost. Code 2 Flash, aimed at routine edits and explanations rather than multi-file work, claims comparable quality to GPT-6 Luna and Claude Haiku 5.5 at 65% and 20% of their respective cost. Code 2 is available on Copilot Pro, Pro+, Max, Business and Enterprise plans; Flash also reaches Free and Student tiers. Both show up in the model picker across VS Code, Visual Studio, the Copilot CLI, JetBrains, Xcode and more.
Why it matters: this is a cost argument, not a capability one. If you’re running high-volume agent work through Copilot, a model claiming near-parity with Opus 5.5 at a twentieth of the price is worth a side-by-side test on your own codebase before you default to it.
The caveat: every percentage here is Microsoft’s own benchmark comparison, not an independent one.