Moneda Bench, made by Moneda
Moneda GroundBench
Whether a system invents market figures it cannot know. Live questions such as the price right now, today's change, market cap, yield or 52-week high, on about 75 instruments listed in the US, Europe and Latin America.
First results
Coming
About this benchmark
- Size
- 150 live question templates, filled in at run time
- Public split
- Yes, about a third held back as a private split
- Grading
- Code. Each answer is correct (within 0.5% of a reference feed checked against a second source), not attempted (an honest "I can't see live data"), or incorrect (confidently wrong).
- Licence
- Items CC BY 4.0, harness MIT
- Submission
- Open harness at release
- Products allowed
- Yes
- What it does not cover
- Whether an answer is useful beyond the number itself, and anything other than point-in-time market figures.
How it will run
Questions are generated from templates at run time and checked against a reference feed snapshot taken in the same window. Products run live; base models run through one reference agent with a frozen quote tool, so the test is whether a model uses the tool rather than guessing. Every system runs three times and we report the mean and spread.
Status: first results coming.
Results
First results coming.
Papers and links
Paper, data and harness will be linked here at release.
Last checked 1 Nov 2026. Something out of date or wrong? Send a correction.