benchgap
Other · model

MAI Code 2 benchmark scores

As of 2026-11-18, MAI Code 2 (Other) has measured scores on 2 benchmarks and estimated scores on 18 more.

BenchmarkScoreSource
AA-SciCode58.5%estimated ± 3.9 pp, medium confidence
SciCode61.0%estimated ± 7.6 pp, low confidence
Terminal-Bench 2.184.6%measured
AA Coding Index77.1%estimated ± 2.5 pp, high confidence
SWE-bench Verified89.6%estimated ± 5.4 pp, medium confidence
SWE-bench Pro82.8%estimated ± 7.7 pp, medium confidence
AA Terminal-Bench 2.182.4%estimated ± 2.6 pp, high confidence
CursorBench 3.268.0%estimated ± 6.4 pp, low confidence
Terminal-Bench 3.017.8%estimated ± 9.1 pp, low confidence
FrontierCode 1.1 Main51.5%estimated ± 3.5 pp, low confidence
Terminal-Bench 4.121.6%estimated ± 10.9 pp, low confidence
Terminal-Bench 2.1 (Vals AI)71.1%estimated ± 7.3 pp, medium confidence
Bug Hunt Bench35.5%estimated ± 10.4 pp, low confidence
DeepSWE73.2%measured
PostTrainBench v1.141.3%estimated ± 4.5 pp, high confidence
VulcanBench Frontier v489.3%estimated ± 5.0 pp, high confidence
AA Terminal-Bench 4.08.2%estimated ± 11.0 pp, low confidence
NL2Repo62.2%estimated ± 6.0 pp, medium confidence
ScarfBench68.2%estimated ± 3.8 pp, medium confidence
SWE-Atlas Codebase QnA61.3%estimated ± 4.8 pp, high confidence