FinCUABuildBench (Tool-Augmented Prompting): leaderboard
FinCUABuildBench (Yang et al., 2026) measures whether AI systems can build reliable evaluation tasks for financial computer-use agents: for each of 576 construction requests the system must produce a grounded, executable and replayable episode with a runtime event and a hidden verifier. Qualified task rate in percent, counting only episodes that pass every hard gate, when the model writes the episodes with access to finance data and application tools; higher is better.
Metric: Qualified task rate (%; share of the 576 preregistered construction slots, 24 financial workflow categories x 8 WorkflowCores x 3 runtime-dynamic profiles, whose generated computer-use episode passes every hard gate: judge acceptance, source and state traceability, execution and replay, verifier pass, valid runtime event, target-capability match and non-duplication, plus every mandatory safety constraint; mean over the preregistered construction runs, macro-averaged over WorkflowCores; the model writes each episode with access to the finance data and application tools). Source: arxiv.org. Saturation forecast: Around 2030. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiniMax-M3 | 17.5 | #265 |
| 2 | Qwen 3.5 397B A17B | 12.8 | #242 |
| 3 | Kimi K3 | 10.9 | #53 |
| 4 | MiMo-V2.5-Pro | 10.8 | #274 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=fincuabuildbench-tool-augmented-prompting · How It Works · Data refreshed daily, snapshot 2026-11-12.