FaithJudge - RAGTruth Data-to-Text: leaderboard
Share of the model's 150 business overviews written from structured JSON (RAGTruth data-to-text) that the FaithJudge judge flags as hallucinated.
Metric: Hallucination rate (%) on 150 RAGTruth data-to-text items. Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 35 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 4 | #239 | |
| 2 | Gemini 2.5 Pro | 4.67 | #274 | |
| 3 | Gemini 2.5 Flash | 4.67 | #467 | |
| 4 | GPT-4.1 | 8 | #505 | |
| 5 | GPT-4.5 (Preview) | 8 | #499 | |
| 6 | DeepSeek R1 0528 | 11.33 | #431 | |
| 7 | Claude 3.7 Sonnet (20250219) | 12.67 | #421 | |
| 8 | Gemini 2.0 Flash (001) | 13.33 | #727 | |
| 9 | Claude Opus 4 (20250514) | Thinking | 13.33 | #300 |
| 10 | Grok 3 | 14 | #606 | |
| 11 | GPT-3.5 Turbo (0125) | 15.33 | #3575 | |
| 12 | O3 (2025-04-16) | Medium | 15.33 | #229 |
| 13 | Mistral Small 3 | 18 | #1799 | |
| 14 | Claude Sonnet 4 (20250514) | Thinking | 19.33 | #405 |
| 15 | Qwen 2.5 72B Instruct | 20 | #1119 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=faithjudge-ragtruth-data-to-text · How It Works · Data refreshed daily, snapshot 2026-11-17.