Finance agents
FinGAIA
Chinese-language multi-file, multi-tool, multi-step financial agent workflows.
As of 13 August 2026. Conditional pack. Chinese-language financial agent benchmark. Language is part of the treatment.
FinGAIA tests multi-file, multi-tool, multi-step financial agent work in Chinese. English filing packs are not a substitute. End-to-end success and step-level correctness stay separate.
Recheck terms and language coverage before a commercial run. No Fable 5 or GPT-5.6 cell is cited.
- Scale
- Public questions as a single workbook plus attachments. Pin the hosted revision at acquisition.
- Provenance
- FinGAIA public release. Language and rights must be recorded at acquisition.
- Contamination
- Unknown. Chinese-language public release.
- Cannot establish
- English filing-research product performance, or a vendor moat. Recheck terms before a commercial run.
Metrics, kept separate
- End-to-end task success
- Tool-trace completeness
- Step-level correctness
No cited public figure is installed for this pack. Method, scale, and gates still apply. A missing chart is not a pass rate.
What the figures bound
The pack is listed because agentic finance evaluation is moving toward multi-step flows. A Chinese public set does not score an English research terminal.
No chart is installed without a cited official cell.
What this page does not conclude
- English research-product performance
- GPT-5.6 or Fable 5 on this pack
- A commercial run before terms are approved
- A single financial-agent winner
Cited as calibration on these memos — not a score of those products.
Not attached to a software, advisory, or automation memo yet.