Skip to content

Benchmark

Vals MortgageTax

Scores are percentages.

Models measured
88
Scores recorded
93
Sources
1
IQ weight
2.02

Score distribution

60.6%
Mean
64.1%
Median
11%
Std dev
45.4%
10th percentile
68.8%
90th percentile
3.5%
Min
3.5%72.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.02 for chat models, measured over 88 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 5· Vals AI72.1%
2Claude Fable 5.1· Vals AI70.8%
3Claude Opus 4.7· Vals AI70.3%
4Claude Sonnet 5· Vals AI70%
5Claude Opus 4.8· Vals AI69.9%
6Gemini 3.1 Pro Preview· Vals AI69.4%
7Gemini 3 Pro Preview· Vals AI69.1%
8Claude Fable 5· Vals AI68.9%
9Gemini 2.5 Pro· Vals AI68.9%
10GPT 5.5· Vals AI68.8%
11Gemini 3 Flash· Vals AI68.7%
12Claude Opus 4.5· Vals AI68.7%
13Claude Sonnet 3.7· Vals AI68.7%
14Gemini 3.5 Flash Lite· Vals AI68.7%
15Claude Opus 4.6· Vals AI68.5%
16MiniMax M3· Vals AI68.4%
17GPT 5.4· Vals AI68.3%
18Qwen3.6 27B· Vals AI68.3%
19Gemini 3.5 Flash· Vals AI68.1%
20Gemini 3.1 Flash Lite· Vals AI68%
21Qwen3.6 Plus· Vals AI68%
22Claude Sonnet 4.6· Vals AI67.7%
23Gemini 3.6 Flash· Vals AI67.6%
24Qwen3.5 Flash· Vals AI67.4%
25GPT-5.6 Terra· Vals AI67.3%
26GPT-5.6 Luna· Vals AI67.3%
27GPT-5.6 Sol· Vals AI67.3%
28Gemini 2.5 Pro· Vals AI67.2%
29GPT 5.2· Vals AI67.1%
30GPT-5 mini· Vals AI66.9%
31Claude Sonnet 3.7thinking· Vals AI66.9%
32Gemini 3.7 Flash· Vals AI66.7%
33Kimi K2.5thinking· Vals AI66.5%
34Kimi K3· Vals AI66.3%
35Qwen3.7 Plus· Vals AI66.2%
36Muse Spark 1.1· Vals AI66.1%
37GPT 4.1· Vals AI65.9%
38Kimi K2.6· Vals AI65.8%
39OpenAI o3· Vals AI65.7%
40GPT-4.1 mini· Vals AI65.5%
41GPT 5· Vals AI65.5%
42Muse Spark 1.2· Vals AI65.4%
43Gemini 3.8 Flash· Vals AI65.3%
44Qwen 3.8 27B· Vals AI64.9%
45o4 Mini· Vals AI64.8%
46Grok 4.6· Vals AI64.2%
47Claude 3.5 Sonnet· Vals AI64.1%
48Claude Sonnet 4.5· Vals AI64%
49Inkling FP4· Vals AI64%
50Qwen3.8 Max· Vals AI64%

Showing 1–50 of 93 results.