Laguna M.1/XS.2 Technical Report

Benchmark Model Rank Results
code-generation-on-multipl-eLaguna XS.2#14Accuracy: 58.4
code-generation-on-swe-bench-verifiedLaguna M.1 (pool)#25Accuracy: 74.6
code-generation-on-swe-bench-verifiedLaguna XS.2 (pool)#43Accuracy: 69.9
code-generation-on-swe-bench-multilingualLaguna M.1 (pool)#28Accuracy: 63.1
code-generation-on-swe-bench-multilingualLaguna XS.2 (pool)#32Accuracy: 57.7
code-generation-on-swe-bench-proLaguna M.1#34Resolved: 49.2
code-generation-on-swe-bench-proLaguna XS.2#38Resolved: 46.3
math-word-problem-solving-on-mathLaguna XS.2#57Accuracy: 58.8
multi-task-language-understanding-on-mmlu-proLaguna XS.2#108Accuracy: 53.0
arithmetic-reasoning-on-gsm8kLaguna XS.2#124Accuracy: 84.2