Mahal Systems FinGAIA
Fits general finance research agents in the GAIA style (web, files, code). Borrow the three scenario depths (operational analytics, asset decision, strategic risk) and the capability breakdown per task to size a case set.
What it asks the agent to do
- browse the web to find financial facts
- read and process attached files (CSV, PDF, Word, audio, JSON)
- run code or calculations on financial data
- carry out multi-step analysis across securities, funds, banking or insurance
What to borrow from it
case templatetaxonomybaseline
Caveats
Chinese-language tasks about the Chinese financial market; only part of the 407 tasks is released; the scoring procedure beyond accuracy is not described; web-dependent answers can drift.
Details
| Workloads | finance, web browsing, tool calling, research, documents |
|---|---|
| Interaction | single turn |
| Environment | mixed |
| End state checked | No |
| Tools | web browsing, file readers, code execution |
| Grading | exact match |
| Size | 407 |
| Data | real public |
| Personal data | none |
| License | MIT |
| Harness | own repo |
| Status | active |