DTap-Bench (Benign Tasks) - Finance: leaderboard
Benign financial trading tasks from DTap-Bench (Chen et al., 2026), run by a tool-using agent in simulated MCP environments with no attack applied. Success rate in percent over 200 tasks, averaged over task categories and judged by a rule-based check of the final environment state; higher is better.
Metric: Benign task success rate (%; 200 benign financial trading tasks, no attack applied; a task succeeds when a rule-based check of the final environment state finds the instruction fully completed; mean of its task-category rates). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 97.7 | #172 |
| 2 | GPT-5.4 | 96.6 | #121 |
| 3 | GPT-5.5 | 96.5 | #48 |
| 4 | Gemini 3 Pro | 95.1 | #135 |
| 5 | Claude Sonnet 4.5 | 94.9 | #244 |
| 6 | GPT-5.1 | 94 | #225 |
| 7 | DeepSeek V4 Pro | 92.4 | #181 |
| 8 | GPT-OSS-120B | 12.8 | #699 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=dtap-bench-benign-tasks-finance · How It Works · Data refreshed daily, snapshot 2026-11-12.