Benchmarks / Professional

TaxEval (Vals)

A Vals-created set of questions and responses to tax questions Scores are the values published by VALS AI. Source benchmark version 2.

124models
–Open Index weight
2026-09-01last updated
Vals AIpublisher

TaxEval (Vals)

Higher is better · best effort per model

Chart settings
124 of 124 selected

full table below

#ModelEffortScoreCost / taskHarness
1Muse Spark 1.2xhigh80.4%$0.014–
2Muse Spark 1.1xhigh79.7%$0.015–
3Muse Spark–77.7%$0.0019–
4Sonnet 4.6max77.1%$0.128–
5Fable 5max76.9%$0.220–
6GPT-5.6 Terraxhigh76.2%$0.036–
7GPT-5.6 Lunamax76.2%$0.012–
8Opus 4.6max76.0%$0.071–
9Fable 5.1max76.0%$0.347–
10Grok 3none75.9%$0.017–
11GPT-5.2xhigh75.8%$0.066–
12Kimi K3max75.7%$0.078–
13Grok 4 Fastnone75.7%$0.0013–
14Opus 4.8max75.6%$0.159–
15Sonnet 5max75.6%$0.302–
16GLM 5.3 Flashmax75.6%$0.0015–
17Qwen 3.8 Maxmax75.6%$0.076–
18Inkling Smallxhigh75.5%$0.018–
19Qwen 3.7 Maxmax75.3%$0.028–
20Inklingmax75.3%$0.103–
21Opus 4.7max75.3%$0.125–
22GPT-5 Minihigh75.2%$0.012–
23Opus 5max75.1%$0.114–
24GPT-4.1high75.1%$0.0062–
25GPT-5.5xhigh75.0%$0.060–
26Gemini 3.6 Flashhigh74.9%$0.020–
27GPT-5.1high74.9%$0.0072–
28Opus 4.5high74.9%$0.219–
29o4 minihigh74.8%$0.0071–
30GPT-5.6 Solmax74.8%$0.138–
31Gemini 3.7 Flashhigh74.7%$0.019–
32Qwen 3.6 Plusmax74.7%$0.011–
33Kimi K2.6max74.7%$0.029–
34o3high74.6%$0.013–
35GPT-4ohigh74.5%$0.0058–
36Gemini 3.8 Flashhigh74.4%$0.024–
37Gemini 3.5 Flashhigh74.4%$0.038–
38o1high74.3%$0.122–
39Kimi K2.5max74.2%$0.014–
40Grok 4.20xhigh74.1%$0.017–
41Sonnet 3.7medium74.0%$0.051–
42Qwen 3 Maxmax74.0%$0.0067–
43GPT-5.4xhigh74.0%$0.157–
44Gemini 3 Flashhigh73.9%$0.0058–
45MiMo V2.5 Proxhigh73.8%$0.0025–
46Opus 4.1high73.7%$0.093–
47GPT-5high73.4%$0.054–
48GLM 5.2max73.3%$0.015–
49Sonnet 4.5high73.3%$0.041–
50Grok 4.1 Fasthigh73.1%$0.0013–
51Nemotron 3 Ultra 550B A55Bmax73.1%––
52Mistral Large 3none73.1%$0.0016–
53DeepSeek V4 Pro 0813max73.1%$0.046–
54Grok 3 Minilow73.0%$0.0011–
55Gemini 2.5 Prohigh72.9%$0.011–
56Gemini 3.1 Prohigh72.9%$0.031–
57Gemini 2.5 Flashmedium72.7%$0.0031–
58MiniMax M3max72.7%$0.0073–
59Gemini 3.5 Flash-Litehigh72.6%$0.0073–
60Gemini 3 Prohigh72.6%$0.034–
61GLM 4.5xhigh72.4%$0.0096–
62GLM 5.3max72.4%$0.032–
63DeepSeek R1xhigh72.3%$0.012–
64Qwen 3.5 Flashmax72.2%$0.0031–
65DeepSeek V4max72.1%$0.033–
66Sonnet 4high72.0%$0.029–
67GPT-4.1 Minihigh71.9%$0.0012–
68Opus 4medium71.9%$0.037–
69MiMo V2.5xhigh71.8%$0.0007–
70Gemini 3.1 Flash-Litehigh71.8%$0.0010–
71Kimi K2medium71.7%$0.0091–
72Grok 4.5high71.7%$0.019–
73gpt-oss-120bhigh71.6%$0.0030–
74Qwen 3.6 27Bmax71.3%$0.030–
75GPT-5.4 Minixhigh71.2%$0.0068–
76GLM 5.1max71.2%$0.017–
77DeepSeek V3none71.1%$0.0007–
78Grok 4.6high71.1%$0.017–
79Qwen 3.8 27Bxhigh70.8%$0.046–
80Grok 4.3xhigh70.8%$0.014–
81DeepSeek V4 Flash 0731high70.7%$0.0098–
82Ling 3.0 Flash–70.6%$0.0007–
83Qwen 3 235B–70.6%$0.0046–
84Mistral Medium 3.1none70.3%$0.0014–
85Claude 3.5 Sonnet Latest–70.2%$0.0060–
86GLM 5high70.0%$0.022–
87Gemini 2.0 Flashmedium69.8%$0.0008–
88o3 minihigh69.4%$0.018–
89GLM 4.7none68.8%$0.014–
90Gemini 2.0 Pronone68.2%$0.0040–
91MiniMax M2.5high68.2%$0.0086–
92DeepSeek V3.2high68.2%$0.0024–
93Mistral Medium 3.5high68.0%$0.079–
94Haiku 4.5xhigh67.5%$0.027–
95GPT-5.4 Nanohigh67.4%$0.0006–
96GPT-5 Nanohigh67.4%$0.0049–
97Grok 2none67.0%$0.0069–
98Llama 4 Maverickmedium66.6%$0.0005–
99MiniMax M2.7high66.6%$0.0035–
100MiniMax M2.1–66.4%$0.0091–
101GLM 4.6none66.2%$0.0093–
102Gemini 2.5 Flash Litenone66.2%$0.0005–
103Grok 4high65.1%$0.078–
104Mistral Large 11–63.8%$0.0041–
105Llama 3.3none63.7%––
106gpt-oss-20bhigh63.7%$0.0023–
107Magistral Medium 1.2medium61.9%$0.025–
108Command Anone61.4%$0.0065–
109Jamba 1.6 Largenone60.9%$0.0052–
110Llama 3.1max60.9%$0.0019–
111GPT-4.1 Nanohigh60.8%$0.0003–
112GPT-4o minihigh60.5%$0.0003–
113Magistral Small 1.2–60.3%$0.0039–
114Gemini 1.5 Pronone59.5%$0.0025–
115Laguna XS.2–59.0%$0.0003–
116Mistral Small 3.1none58.3%$0.0002–
117Jamba 1.5 Largenone58.2%$0.0054–
118Haiku 3.5medium57.4%$0.0013–
119Llama 4 Scoutmedium55.2%$0.0004–
120Mistral Small 02–49.1%$0.0004–
121Gemini 1.5 Flashnone48.2%$0.0002–
122Jamba 1.6 Mininone44.6%$0.0003–
123Jamba 1.5 Mininone41.9%$0.0003–
124Laguna M.1–1.6%––