RPCBench - Domain - Yelp Local - Evidence Faithfulness: leaderboard
Metric: Faithfulness Index (%). Source: github.com. Saturation forecast: Around December 2026. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 397B A17B | 90.08 |
| 2 | Qwen 3.5 Plus | 89.96 |
| 3 | GPT-5.5 | 89.77 |
| 4 | Qwen 3.5 35B A3B | 89.69 |
| 5 | Qwen 3.5 122B A10B | 89.66 |
| 6 | Gemini 3.1 Pro (Preview) | 87.18 |
| 7 | DeepSeek V4 Flash | 83.7 |
| 8 | DeepSeek V4 Pro | 83.44 |
| 9 | Claude Sonnet 4.6 | 78.93 |
| 10 | Llama 3.1 70B Instruct | 34.62 |
| 11 | Llama 3.1 8B Instruct | 28.55 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-domain-yelp-local-evidence-faithfulness · How It Works · Data refreshed daily, snapshot 2026-09-24.