Deception Effectiveness (Lechmazur): leaderboard
Measures LLM capability to generate persuasive disinformation: models craft misleading arguments for incorrect answers to fact-based questions.
Metric: Deception Score. Source: github.com. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet | 1.1 | #350 |
| 2 | Mistral Large 2 (Jul) | 1.09 | #484 |
| 3 | O1 Preview | 1.03 | #193 |
| 4 | Grok 2 | 0.96 | #452 |
| 5 | Gemini 1.5 Pro (Sept) | 0.94 | #448 |
| 6 | Llama 3.1 405B | 0.78 | #466 |
| 7 | Llama 3.1 70B | 0.71 | #592 |
| 8 | O1 Mini | 0.67 | #323 |
| 9 | Claude 3 Haiku | 0.66 | #803 |
| 10 | Claude 3 Opus | 0.65 | #571 |
| 11 | DeepSeek V2.5 | 0.61 | #489 |
| 12 | Gemini 1.5 Flash | 0.61 | #618 |
| 13 | GPT-4o | 0.6 | #343 |
| 14 | GPT-4 Turbo | 0.56 | #404 |
| 15 | Gemma 2 27B | 0.52 | #806 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=deception-effectiveness-lechmazur · How It Works · Data refreshed daily, snapshot 2026-10-20.