Atlas

Benchmarks

Financial Coaching Bench - Elicitation

Financial Coaching Bench's Elicitation panel: the share of the scenarios' hidden fact-sheet facts (income, debts, deadlines, savings, household) that the coach asked the simulated client for, as Jev 1.13, a separate classification model, grades each fact asked for, assumed or never raised. The page gives neither the number of facts nor whether the share is pooled over facts or averaged over scenarios. Diagnostic only.

Models
26
Top score
69.1
Leader
Muse Spark 1.3
Updated
—
Resolution
—
Saturation
—

No signal or difficulty estimate. Held out of the fit by design.

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

Financial Coaching Bench - Elicitation leaderboard

#ModelSettingScoreSource
1Muse Spark 1.3max
69.1
2Muse Spark 1.2x-high
68.2
3Muse Wondermax
67.0
4Muse Wonder Fullmax
66.0
5Gemini 4 Argonhigh
63.8
6Claude Sonnet 5.5max
63.1
7GPT-6 Astramax
62.9
8GPT-5.6 Solmax
61.9
9GPT-5.6 Terramax
61.1
10GPT-6.1 Solmax
59.5
11MiMo-V2.6-Pronot reported
59.1
12Claude Opus 5.5max
58.5
13DeepSeek V4.1 Flashhigh
57.7
14GPT-6 Solmax
55.9
15GPT-5.6 Lunamax
54.4
16GLM-5.3max
54.2
17Gemini 3.8 Flashhigh
54.0
18MiniMax M3—
53.8
19GPT-6 Lunamax
53.3
20Claude Fable 5.1max
53.2
21Grok 4.7x-high
51.7
22Gemini 3.5 Flash-Litehigh
51.3
23Gemini 3.1 Pro Previewhigh
49.8
24Gemini 3.5 Flashhigh
49.7
25Kimi K3max
49.6
26Gemini 3.6 Flashhigh
48.7

Related benchmarks

Loading Atlas data…