LLM Stats (OpenAI-MRCR: 2 needle 128k): leaderboard
OpenAI Multi-Round Coreference Resolution with 2 needles at 128K context, testing precise retrieval of 2 target items in long conversations.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 95.2 | #98 |
| 2 | MiniMax M1 40k | 76.1 | #683 |
| 3 | MiniMax M1 80k | 73.4 | #432 |
| 4 | GPT-4.1 | 57.2 | #248 |
| 5 | GPT-4.1 Mini | 47.2 | #355 |
| 6 | GPT-4.5 | 38.5 | #274 |
| 7 | GPT-4.1 Nano | 36.6 | #731 |
| 8 | GPT-4o | 31.9 | #343 |
| 9 | O3 Mini | 18.7 | #272 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-openai-mrcr-2-needle-128k · How It Works · Data refreshed daily, snapshot 2026-10-20.