Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics

Domain specific AI benchmarks stories

Oct 2, 2026

AI models now outperform licensed CPAs on speed and accuracy in accounting tasks, Mercor study finds

A Mercor study found that current AI models surpass licensed CPAs with an average of 5.5 years of experience in both speed and accuracy on structured bookkeeping tasks, a dramatic improvement from 18 months ago when AI lagged behind. However, no AI model can fully complete all tasks on the more demanding APEX Benchmark without human oversight.

Oct 2, 2026·1 source
00
Aug 15, 2026

New benchmark shows AI models struggle with visual perception, none reach 60% accuracy

Moonshot AI's PerceptionBench reveals that leading multimodal AI models, including GPT-5.6 Sol, perform poorly at basic visual perception tasks when separated from logical reasoning, with no frontier model achieving 60% accuracy. The benchmark demonstrates that many errors attributed to reasoning actually occur during the image-reading stage.

Aug 15, 2026·1 source
00

Top claims

  • ▪A study by Mercor found that artificial intelligence models are faster, significantly cheaper, and solve structured bookkeeping tasks almost flawlessly, outperforming 12 licensed CPAs who averaged 37 percent accuracy.
  • ▪Mercor noted that the tasks in its APEX Accounting Benchmark study specifically tested areas where artificial intelligence excels, such as locating details and following precise instructions
  • ▪Mercor concluded that artificial intelligence models cannot close financial books without human oversight.

Subtopics

AI research & benchmarks2AI1AI labor & economic disruption1AI safety benchmarks1Future of work1Model evaluation methodology1Multimodal models1

Related timelines

Congress

110 stories

Crypto hacks

100 stories

Ebola outbreak

85 stories

Iran War

209 stories

Russia-Ukraine war

136 stories