0

TaxEval v2

Frontier

A Vals-created set of questions and responses to tax questions

Publisher
Vals AI
Domain
Finance
Published
Jun 2026

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
Vals AI
Attribution policy →

Top score 79.7% by Muse Spark 1.1 - 100 models reporting (51 frontier)

Score history

92
25%44%63%81%100%Feb 24Aug 24Feb 25Aug 25Feb 26Mistral Small (Feb '24)GPT-4o-miniGPT-4oGrok 3Claude Opus 4.6Muse Spark 1.1

Top models

100
TaxEval v2Bar chart with 21 bars. Highest value: Muse Spark 1.1 at 79.7.
21 models

FAQ

What is TaxEval v2?
A Vals-created set of questions and responses to tax questions
What is the current top score on TaxEval v2?
The top reported score is 79.7% by Muse Spark 1.1, across 100 models reporting (51 from frontier labs).