Code Generation on MBPP (MBPP Score) benchmark leaderboard · SOTA2 Research

Code Generation on MBPP (MBPP Score)

35.6MBPP ScoreWIDEN

Evaluation Results

Method	Links
WIDEN2026.04		35.6
DARE Ties2026.04		35.2
Precise Fusion2026.04		35.2
Ties Merging2026.04		34.6
DARE Task Arithmetic2026.04		34.4
Task Arithmetic2026.04		33.8
WIDEN2026.04		33.2
Precise Fusion2026.04		32.4
Instruction Tuned2026.04		32.2
Avg Baseline2026.04		31.6
Precise Fusion2026.04		31
Precise Fusion2026.04		30.6
WIDEN2026.04		30.4
Avg Baseline2026.04		30.2
Avg Baseline2026.04		30.2
WIDEN2026.04		29.4
Code2026.04		29.2
Ties Merging2026.04		29.2
Avg Baseline2026.04		29
DARE Task Arithmetic2026.04		28.6
Task Arithmetic2026.04		28.6
Base2026.04		28.2
DARE Ties2026.04		27.4
Ties Merging2026.04		26.8
DARE Ties2026.04		23.8
Ties Merging2026.04		23.6
Math2026.04		23.4
DARE Ties2026.04		21.2
Task Arithmetic2026.04		20.2
DARE Task Arithmetic2026.04		19.6
Task Arithmetic2026.04		19.6
DARE Task Arithmetic2026.04		19