EdWealth announced the launch of MoneyBench, a benchmark designed to assess how AI systems answer personal‑finance questions. The July round compared Ed, the company’s own AI coach, with OpenAI’s ChatGPT and Google’s Gemini.
Ed secured first place, winning 62.3% of the questions. Gemini captured 20.8% and ChatGPT 17.0%. On a five‑point usefulness scale, Ed averaged 4.24, outperforming ChatGPT’s 3.62 and Gemini’s 3.47.
Answers were evaluated by Claude, an Anthropic AI that does not belong to any of the three tested models. Claude cross‑checked each response against live data sources and disqualified any answer containing a false fact. After Claude’s assessment, Ed’s score was reduced by 5.3 points, but the model still finished ahead of its rivals.
The benchmark follows a May round in which Ed placed third after data‑retrieval shortcomings prompted a rebuild of the system. In July, Ed was tested in its production configuration, while ChatGPT ran on GPT‑5.6 Sol at Pro effort and Gemini on version 3.6 Flash with default settings.
Ed’s differentiators include access to more than 120 financial data and analysis tools for live filings, prices and holdings, as well as a base of 11,566 parameters drawn from the IRS, SSA, CMS and state authorities across 51 U.S. jurisdictions. Account connections are read‑only.
An Intuit Credit Karma survey of 1,019 U.S. adults conducted Aug. 7‑14, 2025 found that 52% of respondents who acted on AI‑generated financial advice reported making a poor decision.
Founder Allen Ng said the benchmark puts the product on public trial and emphasized that useful, fact‑correct advice is the core test for financial AI. Ed is available via edwealth.ai, the App Store and Google Play.
