LLM-as-a-Reviewer: leaderboard
Metric: Low (NeurIPS 2022) (self-reported). Source: benchmarklist.com. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Haiku 4.5 | 99.7 |
| 2 | GPT-5.4 Mini | 99.7 |
| 3 | GPT-5.4 | 97.3 |
| 4 | GPT-5 Mini | 79.3 |
| 5 | GPT-4.1 Mini | 60.7 |
| 6 | Gemini 3.1 Flash Lite | 48.3 |
| 7 | Qwen 3 235B A22B | 48.3 |
| 8 | Gemini 2.5 Flash | 46.7 |
| 9 | Gemini 3 Flash | 44.3 |
| 10 | Qwen 3.5 9B | 29.3 |
| 11 | GPT-4o Mini | 28 |
| 12 | Llama 3.3 70B | 6.7 |
Interactive version: theaggregate.ai/benchmark?slug=llm-as-a-reviewer · How It Works · Data refreshed daily, snapshot 2026-09-05.