ModelsMulti-source11:22Chinese contentAPEX-Accounting基准:评估九个前端模型会计能力想看看AI到底能不能做会计?新出的APEX-Accounting基准拿160道真实会计题考了九个模型,最强Claude Fable-5才56%,离实际干活还差得远。#APEX-Accounting#Claude-Fable-5#Muse-Spark-1.1#会计自动化2 个信源在谈EventaarXiv cs.AI@Julien Benchek et al. (11)3 sourcesSourceRead laterReadWorth followingUsefulFollow APEX-Accounting