EnterpriseRAG-Bench: leaderboard
Enterprise retrieval-augmented generation benchmark over company-internal knowledge questions, scoring RAG systems on retrieval, answer quality, and groundedness.
Metric: Overall Score. Source: huggingface.co. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Mixedbread (+ Opus 5) | 86.58 |
| 2 | Prism (aurait.ai) | 80.46 |
| 3 | metor.com | 80.34 |
| 4 | ZNV AgentCube | 80.26 |
| 5 | Skyller AI | 79.3 |
| 6 | CDL (Causal Dynamics Lab) | 78.95 |
| 7 | Troml | 76.79 |
| 8 | AHOY Labs | 74.95 |
| 9 | OpenClaw | 68.22 |
| 10 | SovraRAG.ch | 65.61 |
Interactive version: theaggregate.ai/benchmark?slug=enterpriserag-bench · How It Works · Data refreshed daily, snapshot 2026-09-19.