Proprietary

Vals Index

Updated 5/6/2026

Benchmark consisting of a weighted performance across finance and coding tasks. Showing the potential impact that LLMs can have on the economy.

As of May 6, 2026, Claude Opus 4.7 ranks first on the Vals Index v1 with 72.21%, followed by GPT 5.5 (71.00%) and GPT 5.4 (xhigh) at 66.13%.

Vals Index
ACCURACY

Vals Index leaderboard

Rank Model Accuracy Cost / Test Input / Output Cost Duration
1 Claude Opus 4.7 72.21% $5.51 $5 / $25 12m01s
2 GPT 5.5 71.00% $4.80 $5 / $30 18m33s
3 GPT 5.4 (xhigh) 66.13% $2.79 $2.5 / $15 24m44s
4 Claude Opus 4.6 (Thinking) 66.00% $2.50 $5 / $25 10m05s
5 Claude Sonnet 4.6 65.55% $1.95 $3 / $15 10m38s
6 GPT 5.2 64.52% $4.21 $1.75 / $14 23m32s
7 DeepSeek V4 62.22% $0.76 $1.32 / $3.96 18m30s
8 Kimi K2.6 59.33% $0.59 $0.95 / $4 15m58s
9 Gemini 3.1 Pro Preview (02/26) 58.11% $1.26 $2 / $12 7m59s
10 GLM 5.1 57.92% $0.80 $1 / $3.2 14m11s
11 Muse Spark 56.58% N/A N/A 8m33s
12 GPT 5.4 Mini 55.71% $0.66 $0.75 / $4.5 18m45s
13 GLM 5 55.09% $8.33 $1 / $3.2 52m00s
14 Claude Opus 4.5 (Thinking) 54.60% $7.44 $5 / $25 11m38s
15 GPT 5.1 53.65% $0.87 $1.25 / $10 12m54s
16 Qwen 3.6 Plus 53.34% $1.34 $0.5 / $3 13m12s
17 Claude Sonnet 4.5 (Thinking) 53.24% $1.94 $3 / $15 14m24s
18 Gemini 3 Flash (12/25) 53.16% $0.36 $0.5 / $3 8m12s
19 Grok 4.3 52.52% $0.63 $1.25 / $2.5 11m34s
20 Gemini 3 Pro (11/25) 52.12% $1.70 $2 / $12 38m04s
21 GPT 5.4 Nano 51.88% $0.33 $0.2 / $1.25 21m02s
22 Kimi K2.5 51.70% $0.29 $0.6 / $3 13m27s
23 Qwen 3.5 Plus 50.81% $1.07 $0.4 / $2.4 19m15s
24 MiniMax-M2.7 49.85% $0.72 $0.3 / $1.2 14m51s
25 GPT 5 48.99% $0.58 $1.25 / $10 14m19s
26 Qwen 3.6 27B 48.70% $1.29 $0.6 / $3.6 44m23s
27 Claude Haiku 4.5 (Thinking) 47.22% $0.50 $1 / $5 6m05s
28 Grok 4.20 (Reasoning) 46.07% $0.40 $2 / $6 2m19s
29 GPT 5 Mini 44.47% $0.12 $0.25 / $2 8m54s
30 Grok 4 Fast (Reasoning) 42.64% $0.06 $0.2 / $0.5 6m56s
31 Grok 4.1 Fast (Reasoning) 42.50% $0.07 $0.2 / $0.5 4m14s
32 Gemini 3.1 Flash Lite Preview 40.77% $0.16 $0.25 / $1.5 5m01s
33 Gemini 2.5 Pro 39.32% $0.51 $1.25 / $10 16m39s
34 Mistral Small 4 29.07% $1.39 $0.15 / $0.6 9m14s
35 DeepSeek V3.2 (Thinking) 26.34% $0.73 $0.56 / $1.68 24m52s

Results

Industry Average Accuracy Comparison

Methodology

This archived version uses the original Vals Index finance component, before the Finance Agent v2 index subset replaced it in v1.1.

Finance (weight: 2.0): ~$2T contribution to U.S. GDP

Coding (weight: 1.4): ~$1.4T contribution to U.S. GDP

Coding = 0.25 * SWE_Bench + 0.25 * TBench + 0.5 * VibeCodeBench
Vals_Index = (2.0 * AVG(CorpFin, FinanceAgent) + 1.4 * Coding) / 3.4