Skip to DrivenBench results

August 25, 2026

DrivenBench

1.0

Compare AI models for investment agents across capability, cost, and latency, using real-world investment workflows and tasks.

Why DrivenBench?
DrivenBench score

Capability score vs. total benchmark cost upper bound (log scale). The solid line shows the observed Pareto frontier.

Cost upper bound (USD, log scale)

Leaderboard

DrivenBench capability scores, capability pass counts, total benchmark cost upper bounds, and median and p90 latency
RankModelScore
=193.9%
=193.9%
390.9%
489.7%
589.1%
687.3%
785.5%
881.8%
979.4%
1077.6%
1176.4%

Choose the model that fits your investment workflow, then put it to work in Driven.

Try Driven Free
Driven wordmark