REAL: leaderboard
Real Estate Agent LLM Evaluation suite for housing transactions and services, covering memory, comprehension, reasoning, and hallucination across Beijing housing-service scenarios.
Metric: Average Accuracy (%). Source: arxiv.org. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 50.64 | #325 |
| 2 | GPT-4o (2024-08-06) | 50.45 | #336 |
| 3 | Qwen 2 72B Instruct | 48.62 | #576 |
| 4 | Qwen 2.5 72B Instruct | 48.25 | #454 |
| 5 | GPT-4 Turbo | 42.5 | #404 |
| 6 | Yi 1.5 34B Chat | 40.07 | #800 |
| 7 | GLM-4 9B Chat | 38.98 | #916 |
| 8 | Baichuan2-13B-Chat | 36.6 | #1255 |
| 9 | DeepSeek-V2-Lite-Chat | 35.77 | #1210 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=real · How It Works · Data refreshed daily, snapshot 2026-10-20.