Vending-Bench Arena - Round 3: leaderboard
Metric: Final money balance ($, mean over runs). Source: andonlabs.com. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash | 3436.09 |
| 2 | Claude Haiku 4.5 | 1700.51 |
| 3 | Grok 4.1 Fast | 723.11 |
| 4 | Gemini 2.5 Flash | 280.75 |
| 5 | GPT-5 Mini | 153.66 |
Interactive version: theaggregate.ai/benchmark?slug=vending-bench-arena-round-3 · How It Works · Data refreshed daily, snapshot 2026-09-19.