Vals AI MortgageTax — leaderboard

Metric: Accuracy (%). Source: www.vals.ai. 87 models tracked.

Top models

#ModelScore
1Claude Opus 4.770.27
2Claude Sonnet 570.03
3Claude Opus 4.869.91
4Gemini 3.1 Pro (Preview)69.4
5Gemini 3 Pro (Preview)69.08
6Gemini 2.5 Pro68.92
7Claude Fable 568.92
8GPT-5.568.76
9Gemini 3 Flash (Preview)68.72
10Claude 3.7 Sonnet (20250219)68.68
11Claude Opus 4.5 (20251101)68.68
12Claude Opus 4.6 (Thinking)68.52
13MiniMax-M368.36
14GPT-5.4 (2026-03-05)68.32
15Qwen 3.6 27B68.28

Interactive version: theaggregate.ai/benchmark?slug=vals-ai-mortgagetax · How the rankings work · Data refreshed daily, snapshot 2026-07-22.