LUNAR - Curated Context: leaderboard
Metric: Personalization score (1-5; mean of coverage and depth, GPT-5.1 judge). Source: arxiv.org. 19 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K2.6 | 4.07 |
| 2 | Gemini 3 Flash (Preview) | 4.02 |
| 3 | GLM-5.1 | 3.86 |
| 4 | DeepSeek V4 Flash | 3.78 |
| 5 | DeepSeek V4 Pro | 3.66 |
| 6 | MiniMax-M2.7 | 3.57 |
| 7 | GPT-4.1 Mini | 3.39 |
| 8 | Qwen 3 30B A3B 2507 Instruct | 3.37 |
| 9 | Ling-2.6-1T | 3.35 |
| 10 | Qwen 3 14B (Non-reasoning) | 3.27 |
| 11 | Ling-2.6-flash | 3.19 |
| 12 | GLM-4.7 Flash | 3.15 |
| 13 | Qwen 3 8B (Non-reasoning) | 3.12 |
| 14 | Qwen 3 32B (Non-reasoning) | 3.12 |
| 15 | GPT-4o Mini | 3.08 |
Interactive version: theaggregate.ai/benchmark?slug=lunar-curated-context · How It Works · Data refreshed daily, snapshot 2026-09-19.