Atlas

Benchmarks

Financial Coaching Bench

Vals AI's Financial Coaching Bench v1, supported by NextLadder Ventures and advised by the Financial Health Network, tests whether a model can coach a household through a financial problem such as a collections notice, a rent increase or a defaulted loan. In each of 256 of its 269 private scenarios, spread over seven domains (Borrow, Spend, Save, Plan, Life Events & Shocks, Long-Term Wealth & Security, and Emerging & Digital-First Behaviors), the model can ask questions and search the web while Claude Sonnet 4.6 plays the client from a hidden fact sheet, until the coach delivers a plan or reaches the exchange limit. DeepSeek V4.1 Flash grades each practitioner-written checklist item against the transcript (must-have items strictly, path-dependent ones also passing with a justified alternative); the score is the share of items met, averaged over scenarios, and an unfinished scenario (timeout or context overflow) scores zero.

Models
26
Top score
76.2
Leader
Claude Opus 5.5
Updated
—
Resolution
~5 mo
Useful
Saturation
2028–2030
Projected
Signal
32ndpercentile
195 benchmarks from 0.8 to 3.5224 benchmarks from 3.5 to 6.1133 benchmarks from 6.1 to 8.898 benchmarks from 8.8 to 11.566 benchmarks from 11.5 to 14.126 benchmarks from 14.1 to 16.815 benchmarks from 16.8 to 19.59 benchmarks from 19.5 to 22.18 benchmarks from 22.1 to 24.84 benchmarks from 24.8 to 27.55 benchmarks from 27.5 to 30.12 benchmarks from 30.1 to 32.84 benchmarks from 32.8 to 35.45 benchmarks from 35.4 to 38.11 benchmark from 38.1 to 40.82 benchmarks from 43.4 to 46.11 benchmark from 62.1 to 64.8

539th strongest of 798

Signal-to-noise +4.09

Difficulty
28thpercentile
1 benchmark from 5.5 to 14.41 benchmark from 14.4 to 23.22 benchmarks from 32.1 to 41.01 benchmark from 41.0 to 49.82 benchmarks from 49.8 to 58.79 benchmarks from 58.7 to 67.67 benchmarks from 67.6 to 76.418 benchmarks from 76.4 to 85.324 benchmarks from 85.3 to 94.241 benchmarks from 94.2 to 103.031 benchmarks from 103.0 to 111.939 benchmarks from 111.9 to 120.850 benchmarks from 120.8 to 129.676 benchmarks from 129.6 to 138.560 benchmarks from 138.5 to 147.4118 benchmarks from 147.4 to 156.2124 benchmarks from 156.2 to 165.1120 benchmarks from 165.1 to 174.039 benchmarks from 174.0 to 182.820 benchmarks from 182.8 to 191.710 benchmarks from 191.7 to 200.61 benchmark from 200.6 to 209.44 benchmarks from 209.4 to 218.3

576th hardest of 798

Capability needed ≲ 129.2

Organizations
26 dated models · 5 frontier modelsClosed weightsOpen weights
Data: Vals AIatlas.village.dev

Financial Coaching Bench leaderboard

#ModelSettingScoreSource
1Claude Opus 5.5max
76.2
2Muse Spark 1.2x-high
75.7
3Claude Fable 5.1max
75.7
4MiMo-V2.6-Pronot reported
75.4
5Muse Wonder Fullmax
75.0
6Claude Sonnet 5.5max
74.9
7GPT-5.6 Solmax
74.9
8Kimi K3max
74.5
9Grok 4.7x-high
74.1
10GLM-5.3max
74.1
11Muse Wondermax
73.8
12Gemini 4 Argonhigh
73.1
13DeepSeek V4.1 Flashhigh
72.8
14Muse Spark 1.3x-high
71.0
15GPT-5.6 Terramax
71.0
16GPT-5.6 Lunamax
69.3
17MiniMax M3—
69.0
18GPT-6.1 Solmax
68.5
19GPT-6 Astramax
67.5
20Gemini 3.1 Pro Previewhigh
66.7
21Gemini 3.8 Flashhigh
66.3
22Gemini 3.6 Flashhigh
65.8
23GPT-6 Lunamax
65.2
24Gemini 3.5 Flashhigh
63.0
25GPT-6 Solmax
62.6
26Gemini 3.5 Flash-Litehigh
57.5

Related benchmarks

Loading Atlas data…