BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

MortgageTax leaderboard

As of 6 Nov 2026, the top score on MortgageTax is 72.1% by Claude Opus 5 (Anthropic), ahead of Claude Fable 5.1 at 70.8%, across 94 model configurations with a published result (Vals AI, result dated 1 Sep 2026). Scanned tax certificates must be read and their figures extracted correctly, a document-understanding task with exact answers. MortgageTax is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5 holds the top MortgageTax score at 72.1% (Vals AI, 1 Sep 2026).
  • Claude Fable 5.1 is second on MortgageTax at 70.8% (Vals AI, 1 Sep 2026).
  • MortgageTax: Reading tax certificate images and extracting the figures. Run by Vals AI (Vals AI).
  • 94 model configurations have a published MortgageTax result on BenchLeader (BenchLeader, 6 Nov 2026).
  • MortgageTax does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Reading tax certificate images and extracting the figures. Run by Vals AI.

Published by
Vals AI
Category
Multimodal
Index weight
Reference only
Models
94
Data as of
6 Nov 2026

Vals AI (vals.ai).

What the test looks like

Scanned tax certificates must be read and their figures extracted correctly, a document-understanding task with exact answers.

How it is scored

Accuracy, run by Vals AI.

What to keep in mind

Document OCR and reasoning together; only vision models are listed.

Top 15
  1. 1Claude Opus 572.1%
  2. 2Claude Fable 5.170.8%
  3. 3Claude Opus 4.770.3%
  4. 4Claude Sonnet 570.0%
  5. 5Claude Opus 4.869.9%
  6. 6Gemini 3.1 Pro (high)69.4%
  7. 7Gemini 3 Pro (high)69.1%
  8. 8Claude Fable 568.9%
  9. 9Gemini 2.5 Pro68.9%
  10. 10GPT-5.5 (xhigh)68.8%
  11. 11Gemini 3 Flash (high)68.7%
  12. 12Claude Opus 4.568.7%
  13. 13Claude 3.7 Sonnet68.7%
  14. 14Gemini 3.5 Flash Lite (high)68.7%
  15. 15Claude Opus 4.6 (thinking)68.5%
94 of 94
#
1Claude Opus 572.1%65.2$0.0742026-09-01
2Claude Fable 5.170.8%66.4$0.1722026-09-01
3Claude Opus 4.770.3%63.1$0.0842026-09-01
4Claude Sonnet 570.0%55.9$0.0802026-09-01
5Claude Opus 4.869.9%60.0$0.1062026-09-01
6Gemini 3.1 Prohigh69.4%60.1$0.0242026-09-01
7Gemini 3 Prohigh69.1%60.1$0.0412026-09-01
8Claude Fable 568.9%66.5$0.1752026-09-01
9Gemini 2.5 Pro68.9%54.5$0.00482026-09-01
10GPT-5.5xhigh68.8%64.3$0.0712026-09-01
11Gemini 3 Flashhigh68.7%58.3$0.00752026-09-01
12Claude Opus 4.568.7%57.9$0.0922026-09-01
13Claude 3.7 Sonnet68.7%49.9$0.0162026-09-01
14Gemini 3.5 Flash Litehigh68.7%47.9$0.00432026-09-01
15Claude Opus 4.6thinking68.5%62.4$0.0612026-09-01
16MiniMax M368.4%56.4$0.00832026-09-01
17GPT-5.4xhigh68.3%63.2$0.1032026-09-01
18Qwen3.6 27B68.3%48.0$0.0172026-09-01
19Gemini 3.5 Flashhigh68.1%62.5$0.0182026-09-01
20Gemini 3.1 Flash Litehigh68.0%48.5$0.00282026-09-01
21Qwen3.6 Plus68.0%59.1$0.0132026-09-01
22Claude Sonnet 4.667.7%58.2$0.0452026-09-01
23Claude Opus 4.5thinking67.7%59.6$0.1782026-09-01
24Gemini 3.6 Flashhigh67.6%61.2$0.0202026-09-01
25Qwen3.5 Flash67.4%51.8$0.00222026-09-01
26GPT-5.6 Terraxhigh67.3%60.3$0.0232026-09-01
27GPT-5.6 Solmax67.3%65.4$0.0712026-09-01
28GPT-5.6 Lunamax67.3%56.3$0.00482026-09-01
29GPT-5.2xhigh67.1%60.0$0.0502026-09-01
30GPT-5 minihigh66.9%54.6$0.00442026-09-01
31Claude 3.7 Sonnetthinking66.8%51.9$0.0242026-09-01
32Gemini 3.7 Flashhigh66.7%62.2$0.0112026-09-01
33Kimi K2.5thinking66.5%58.7$0.0152026-09-01
34Kimi K366.3%59.1$0.0392026-09-01
35Qwen3.7 Plus66.2%57.7$0.00762026-09-01
36Muse Spark 1.1xhigh66.1%59.5$0.0222026-09-01
37GPT-4.1high65.9%46.6$0.00622026-09-01
38Kimi K2.665.8%55.2$0.0292026-09-01
39o3high65.7%53.8$0.01002026-09-01
40GPT-4.1 minihigh65.5%46.3$0.00282026-09-01
41GPT-5high65.5%57.9$0.0292026-09-01
42Muse Spark 1.2xhigh65.4%61.7$0.0262026-09-01
43Gemini 3.8 Flashhigh65.3%62.0$0.0472026-09-01
44Qwen3.8 27Bxhigh64.9%52.5$0.0122026-09-01
45o4-minihigh64.8%50.3$0.0122026-09-01
46Grok 4.6high64.2%60.5$0.0142026-09-01
47Claude 3.5 Sonnet64.1%47.1$0.0152026-09-01
48Qwen3.8 Maxmax64.0%58.8$0.0272026-09-01
49Claude Sonnet 4.5thinking64.0%52.2$0.0522026-09-01
50Inkling64.0%47.5$0.0562026-09-01
51GPT-5.4 Minixhigh63.5%56.0$0.0452026-09-01
52Gemini 2.5 Flash 09 202562.6%46.8$0.00122026-09-01
53Claude Sonnet 462.5%50.5$0.0162026-09-01
54Inkling Small62.3%52.1$0.0122026-09-01
55Claude Haiku 4.5thinking62.2%45.4$0.0212026-09-01
56Magistral Small 1.262.1%39.9$0.00452026-09-01
57Gemini 2.5 Flash 09 2025thinking61.9%52.8$0.00122026-09-01
58Grok 4.5high61.8%56.9$0.0152026-09-01
59GPT-5.1high61.4%57.7$0.0262026-09-01
60Gemma 4 31Bhigh61.4%––2026-09-01

More multimodal benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on MortgageTax

Cite as: BenchLeader, “MortgageTax leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_mortgage_tax.

MortgageTax: questions

What does MortgageTax measure?
Scanned tax certificates must be read and their figures extracted correctly, a document-understanding task with exact answers. Scores are reported in percent of tasks solved; higher is better.
What is the top score on MortgageTax?
As of 6 Nov 2026, the highest score on MortgageTax is 72.1% by Claude Opus 5 (result dated 1 Sep 2026), ahead of Claude Fable 5.1 at 70.8%. Scores come from Vals AI and are re-read daily.
Which AI model leads MortgageTax?
Claude Opus 5 leads MortgageTax with 72.1% as of 6 Nov 2026 (result dated 1 Sep 2026), ahead of Claude Fable 5.1 at 70.8%.
How many models have MortgageTax results?
94 model configurations have a MortgageTax result on BenchLeader, all taken from Vals AI.
Who runs MortgageTax and how often is it updated?
MortgageTax is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
Does MortgageTax count toward the BenchLeader Index?
No. MortgageTax is shown for reference but left out of the composite index.