Leaderboard, made by Vals AI

TaxEval v2

Hard tax questions, created with financial and tax experts. Scores answer correctness and stepwise reasoning against expert reasoning.

Best reported

80.4

Muse Spark 1.2Meta

accuracy % (v2)

1 Sept 2026

TaxEval v2 v2

accuracy % (v2), higher is better. 5 models.

Updated 1 Sept 2026

  1. Muse Spark 1.280.4
  2. Muse Spark 1.179.7
  3. Muse Spark77.7
  4. Claude Sonnet 4.677.1
  5. Claude Fable 576.9

TaxEval v2 v2

accuracy % (v2), higher is better. 5 models.

Updated 1 Sept 2026

  1. 1Muse Spark 1.280.4
  2. 2Muse Spark 1.179.7
  3. 3Muse Spark77.7
  4. 4Claude Sonnet 4.677.1
  5. 5Claude Fable 576.9
TaxEval v2 leaderboard (accuracy % (v2), higher is better)
1Muse Spark 1.2Meta80.4test set (1,223 questions), archived boardVals.ai TaxEval v2 leaderboard
Official board, reported 1 Sept 2026 · checked 2 Nov 2026
2Muse Spark 1.1Meta79.7test set (1,223 questions), archived boardVals.ai TaxEval v2 leaderboard
Official board, reported 1 Sept 2026 · checked 2 Nov 2026
3Muse SparkMeta77.7test set (1,223 questions), archived boardVals.ai TaxEval v2 leaderboard
Official board, reported 1 Sept 2026 · checked 2 Nov 2026
4Claude Sonnet 4.6Anthropic77.1test set (1,223 questions), archived boardVals.ai TaxEval v2 leaderboard
Official board, reported 1 Sept 2026 · checked 2 Nov 2026
5Claude Fable 5Anthropic76.9test set (1,223 questions), archived boardVals.ai TaxEval v2 leaderboard
Official board, reported 1 Sept 2026 · checked 2 Nov 2026

Full details for TaxEval v2