Skip to content

Benchmark

Vals Tax Agent Bench

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.

Models measured
85
Scores recorded
85
Sources
1
IQ weight
2.00

Score distribution

25.1%
Mean
25.8%
Median
12.8%
Std dev
7%
10th percentile
43%
90th percentile
2.5%
Min
2.5%52.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.00 for chat models, measured over 85 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 4 Hydrogen· Vals AI52.3%
2Gemini 4 Carbon· Vals AI51.8%
3Claude Fable 5.1· Vals AI50.7%
4Gemini 4 Argon· Vals AI49.2%
5Claude Opus 5· Vals AI47.1%
6Claude Opus 5.5· Vals AI44.5%
7DeepSeek V4.1 Flash 1017· Vals AI44.4%
8GLM 5.3· Vals AI43.3%
9Muse Wonder Full· Vals AI43%
10Claude Sonnet 5.5· Vals AI42%
11Qwen4 Max· Vals AI40.4%
12Inkling 2 Small· Vals AI38.9%
13Claude Fable 5· Vals AI38.3%
14Muse Ember· Vals AI38.2%
15Muse Spark 1.3· Vals AI37.9%
16Muse Wonder· Vals AI36.3%
17Grok 4.7· Vals AI35.8%
18Hy 4· Vals AI34.4%
19Grok 4.6· Vals AI33.7%
20Kimi K3· Vals AI33.4%
21Muse Spark 1.2· Vals AI32.8%
22Hy4 preview· Vals AI32.6%
23MiMo V2.6 Pro· Vals AI32.4%
24MiniMax M3.1 Flash· Vals AI32.1%
25Gemini 3.8 Flash· Vals AI32.1%
26Qwen3.8 Max· Vals AI32.1%
27Claude Fable 5.2· Vals AI32.1%
28Claude Haiku 5.5· Vals AI31.9%
29GLM 5.3 Flash· Vals AI31.6%
30GPT-5.6 Sol· Vals AI31.1%
31Ember-1· Vals AI31.1%
32Qwen 3.8 27B· Vals AI30.9%
33Mistral Large 4· Vals AI30.4%
34GPT-5.6 Terra· Vals AI30%
35Claude Opus 4.8· Vals AI30%
36Claude Sonnet 5· Vals AI29.4%
37MiMo V2.6 Flash· Vals AI28.5%
38DeepSeek V4 Flash 0731· Vals AI28.1%
39Step 5 Preview· Vals AI27.3%
40DeepSeek V4 Pro 0813· Vals AI26.8%
41DeepSeek V4.1 Flash· Vals AI26.6%
42Muse Spark Open· Vals AI26.3%
43Claude Sonnet 4.6· Vals AI25.8%
44DeepSeek V4 Pro· Vals AI25.2%
45Kimi K3.1 Pro· Vals AI24.9%
46Grok 4.5· Vals AI24.9%
47GPT-6.2 Sol· Vals AI24.8%
48GPT-5.6 Luna· Vals AI24.8%
49Claude Opus 4.7· Vals AI23.7%
50Hy4 Flash· Vals AI23.3%

Showing 1–50 of 85 results.