Context Arena MRCR (8-needle): leaderboard
OpenAI MRCR benchmark: finds and distinguishes between 8 identical pieces of information in long conversations (8K–1M tokens). Hardest variant, top models score under 50% AUC@1M.
Metric: AUC@1M (%). Source: contextarena.ai. Status: saturation imminent. 52 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Preview) (High) | 49.4 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 48.5 |
| 3 | Gemini 3.1 Pro (Preview) (High) | 47.5 |
| 4 | Gemini 3.1 Pro (Preview) (Low) | 47 |
| 5 | Gemini 3 Flash (Preview) (Medium) | 45.9 |
| 6 | Gemini 3 Pro (Preview) | 39 |
| 7 | Gemini 3 Flash (Preview) | 29.3 |
| 8 | Gemini 2.5 Pro (Preview 06-05) | 28 |
| 9 | Gemini 2.5 Flash (Preview 05-20) | 27 |
| 10 | nova-2-lite-v1 | 25.3 |
| 11 | GPT-5.2 (xHigh) | 19 |
| 12 | GPT-4.1 | 17.5 |
| 13 | GPT-4.1 Mini | 16.7 |
| 14 | Grok 4.1 Fast | 16.5 |
| 15 | GPT-5.2 | 15.5 |
Interactive version: theaggregate.ai/benchmark?slug=context-arena-mrcr-8-needle · How It Works · Data refreshed daily, snapshot 2026-09-05.