Inference
Vals AI

MortgageTax

Evaluating reading and understanding tax certificates as images

ModelsModels with a result.95
Top resultBest result on this test.72.1%Claude Opus 5
Top-3 spreadPoints from first to third.1.8 pts
UpdatedDate the source changed the results.1 Sept 2026

Result and cost

020406080100$0.00003$0.0001$0.0003$0.001$0.003$0.01$0.03$0.1$0.3RESULTUS DOLLARS PER TASK · LOG SCALE

Results

95 results
#ModelResultThe score from the source.CostUS dollars to run one task.TimeTime to run one task.
01Claude Opus 572.1%$0.0748.0 s
02Claude Fable 5.170.8%$0.17215 s
03Claude Opus 4.770.3%$0.08416 s
04Claude Sonnet 570.0%$0.0829 s
05Claude Opus 4.869.9%$0.10621 s
06Gemini 3.1 Pro Preview69.4%$0.02424 s
07Gemini 3 Pro69.1%$0.04126 s
08Claude Fable 568.9%$0.17516 s
09Gemini 2.5 Pro68.9%<$0.0124 s
10GPT-5.568.8%$0.07128 s
11Gemini 3 Flash Preview68.7%<$0.0113 s
12Gemini 3.5 Flash-Lite68.7%<$0.014.9 s
12Claude Opus 4.568.7%$0.09211 s
12Claude 3.7 Sonnet68.7%$0.0165.7 s
15Claude Opus 4.6 (Adaptive)68.5%$0.06127 s
16MiniMax M368.4%<$0.0127 s
17GPT-5.468.3%$0.10345 s
18Qwen3.6-27B68.3%$0.01742 s
19Gemini 3.5 Flash68.1%$0.01811 s
20Gemini 3.1 Flash Lite Preview68.0%<$0.015.4 s
21Qwen3.6 Plus68.0%$0.01378 s
22Claude Sonnet 4.667.7%$0.04530 s
23Claude Opus 4.5 Thinking67.7%$0.17828 s
24Gemini 3.6 Flash67.6%$0.029.2 s
25Qwen3.5 Flash67.4%<$0.0137 s
26GPT-5.6 Terra67.3%$0.0236.3 s
27GPT-5.6 Luna67.3%<$0.0123 s
28GPT-5.6 Sol67.3%$0.07113 s
29Gemini 2.5 Pro Exp 03.2567.2%<$0.018.9 s
30GPT-5.267.1%$0.0571 s
31GPT-5 mini66.9%<$0.0125 s
32Claude 3.7 Sonnet Thinking66.9%$0.02414 s
33Gemini 3.7 Flash66.7%$0.0114.1 s
34Kimi K2.5 Thinking66.5%$0.01580 s
35Qwen3.7 Plus66.2%<$0.0161 s
36Muse Spark 1.166.1%$0.02225 s
37GPT-4.165.9%<$0.015.0 s
38o365.7%<$0.0121 s
39GPT-4.1 mini65.5%<$0.014.6 s
40GPT-565.5%$0.02963 s
41Muse Spark 1.265.4%$0.02645 s
42Gemini 3.8 Flash65.3%$0.04715 s
43Qwen3.8-27B64.9%$0.01280 s
44o4-mini (high)64.8%$0.01213 s
45Grok 4.664.2%$0.0148.5 s
46Claude 3.5 Sonnet64.1%$0.0154.2 s
47Qwen3.8 Max64.0%$0.02747 s
47Inkling64.0%$0.05669 s
47Claude Sonnet 4.5 Thinking64.0%$0.05248 s
50GPT-5.4 mini63.5%$0.0452 min
51Gemini 2.5 Flash (Preview)62.6%<$0.017.3 s
52Claude Sonnet 462.5%$0.0168.4 s
53Inkling-Small62.3%$0.0122 min
54Claude Haiku 4.5 Thinking62.2%$0.02130 s
55Magistral Small62.1%<$0.0111 s
56Gemini 2.5 Flash Preview Thinking61.9%<$0.0118 s
57Grok 4.561.8%$0.0152 min
58GPT-5.161.4%$0.02646 s
58Gemma 4 31B61.4%$042 s
60Mistral Small61.2%<$0.018.3 s
61Claude Opus 4.161.1%$0.08810 s
62Qwen3.5 Plus Thinking60.8%$0.01162 s
63Gemini 2.0 Flash Exp60.4%<$0.012.8 s
64Gemini 2.0 Flash 00159.7%<$0.013.8 s
65MiMo-V2.559.3%<$0.0133 s
66GPT-4o59.2%<$0.017.4 s
67GPT-5.4 nano59.1%<$0.0115 s
68Claude Opus 458.6%$0.08814 s
69Llama4 Maverick Instruct Basic58.5%<$0.012.7 s
70Gemini 2.5 Flash Preview 04.1758.4%<$0.013.6 s
71Llama 4 Scout 17B 16E Instruct57.8%<$0.012.4 s
72Gemini 2.5 Flash Lite Preview Thinking57.6%<$0.0128 s
73Gemini 2.5 Flash Preview 04.17 Thinking57.5%<$0.018.4 s
74Gemini 1.5 Pro 00256.8%<$0.014.2 s
75Claude Opus 4.1 Thinking56.1%$0.13624 s
76Gemini 2.5 Flash Lite Preview55.8%<$0.018.0 s
77GPT-4o mini54.5%<$0.016.1 s
78GPT-5 nano53.6%<$0.0133 s
79GPT-4.1 nano52.8%<$0.013.9 s
80Mistral Large52.1%<$0.019.3 s
81Magistral Medium51.9%$0.0226 s
82Claude Sonnet 4 Thinking49.9%$0.05246 s
83Grok 4.348.3%$0.0127 min
84Grok 4.20 (Reasoning)45.4%$0.02411 s
85Grok 444.5%$0.08665 s
86Gemini 1.5 Flash 00242.8%<$0.012.1 s
87Grok 4.1 Fast (Reasoning)42.6%<$0.0147 s
88Grok 4 Fast (Reasoning)42.1%<$0.0114 s
89Llama 3.2 90B Vision Instruct Turbo38.8%<$0.015.1 s
90Mistral Medium36.4%<$0.017.1 s
91Grok 4.1 Fast Non Reasoning34.3%<$0.018.9 s
92Mistral Medium 3.528.9%$0.02468 s
93Grok 4 Fast Non Reasoning25.3%<$0.014.5 s
94Llama 3.2 11B Vision Instruct Turbo23.3%<$0.012.5 s
95Grok 2 Vision3.5%<$0.010.0 s