Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Multilingual Reasoning on MGSM

49.9Accuracy

U-PaLM

Updated 1mo ago

Evaluation Results

Method	Links
U-PaLM 2022.10		49.9
PaLM 2022.10		45.9

SOTA Paper

U-PaLM

Transcending Scaling Laws with 0.1% Extra Compute

Dataset

MGSM

Follow for update

@wizwand_team Discord

Related Benchmarks

Question Answering on StrategyQA Multilingual Question Answering on TydiQA Zero-shot Commonsense Reasoning on BoolQ, PIQA, HellaSwag, and Winogrande Reading Comprehension on Lambada Closed Book Question Answering on TriviaQA (accuracy)

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs