Benchmarks
347 sources · 1,258 models × 3,038 benchmarks · 158,912 scoresVals 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.
539th strongest of 798 fitted benchmarks
Signal-to-noise +4.09
576th hardest of 798 fitted benchmarks
Capability needed ≲ 129.2
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 | max | 76.2 | |
| 2 | Muse Spark 1.2 | x-high | 75.7 | |
| 3 | Claude Fable 5.1 | max | 75.7 | |
| 4 | MiMo-V2.6-Pro | not reported | 75.4 | |
| 5 | Muse Wonder Full | max | 75.0 | |
| 6 | Claude Sonnet 5.5 | max | 74.9 | |
| 7 | GPT-5.6 Sol | max | 74.9 | |
| 8 | Kimi K3 | max | 74.5 | |
| 9 | Grok 4.7 | x-high | 74.1 | |
| 10 | GLM-5.3 | max | 74.1 | |
| 11 | Muse Wonder | max | 73.8 | |
| 12 | Gemini 4 Argon | high | 73.1 | |
| 13 | DeepSeek V4.1 Flash | high | 72.8 | |
| 14 | Muse Spark 1.3 | x-high | 71.0 | |
| 15 | GPT-5.6 Terra | max | 71.0 | |
| 16 | GPT-5.6 Luna | max | 69.3 | |
| 17 | MiniMax M3 | — | 69.0 | |
| 18 | GPT-6.1 Sol | max | 68.5 | |
| 19 | GPT-6 Astra | max | 67.5 | |
| 20 | Gemini 3.1 Pro Preview | high | 66.7 | |
| 21 | Gemini 3.8 Flash | high | 66.3 | |
| 22 | Gemini 3.6 Flash | high | 65.8 | |
| 23 | GPT-6 Luna | max | 65.2 | |
| 24 | Gemini 3.5 Flash | high | 63.0 | |
| 25 | GPT-6 Sol | max | 62.6 | |
| 26 | Gemini 3.5 Flash-Lite | high | 57.5 |
Benchmarks outside its family whose model rankings track it most reliably (Spearman ρ over at least 10 shared models).