FraudBench: leaderboard
Artificial Intelligence (AI)-generated images have become increasingly realistic and readily adaptable to concrete real-world claims, creating new challenges for verifying visual evidence. A concrete emerging risk is AI-generated refund fraud, in which manipulated or synthetic images are used to support claims about damaged products, poor delivery conditions, or service-related defects. Existing A
Metric: GPT Image 2 TPR (self-reported). Source: benchmarklist.com. Saturation forecast: Not forecast. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 15.8 | #99 |
| 2 | Qwen 3.6 35B A3B | 14.4 | #226 |
| 3 | Grok 4.1 Fast (Reasoning) | 10.5 | #217 (Grok 4.1 Fast) |
| 4 | Qwen 3.6 Plus | 7.7 | #120 |
| 5 | Grok 4.20 (Reasoning) | 6.2 | #175 (Grok 4.20) |
| 6 | GPT-5.4 Mini | 4 | #212 |
| 7 | Kimi K2.6 | 2.6 | #104 |
| 8 | Qwen 3.5 Omni Plus | 0.8 | #84 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=fraudbench · How It Works · Data refreshed daily, snapshot 2026-10-20.