MAI-Code-1-Flash | SWE-bench Multilingual — GitHub Copilot production harness

Microsoft AI's MAI-Code-1-Flash evaluation results on SWE-bench Multilingual — GitHub Copilot production harness

Score: 65.5

Reasoning level: adaptive thinking

Source

MAI-Code-1-Flash (Registry No. 70002)

Microsoft AI

SWE-bench Multilingual — GitHub Copilot production harness