OpenAI-MRCR: 2 needle 1M
Multi-Round Co-reference Resolution benchmark that tests an LLM's ability to distinguish between multiple similar needles hidden in long conversations. Models must reproduce specific instances of content (e.g., 'Return the 2nd poem about tapirs') from multi-turn synthetic conversations, requiring reasoning about context, ordering, and subtle differences between similar outputs.
MiniMax M1 40K from MiniMax currently leads the OpenAI-MRCR: 2 needle 1M leaderboard with a score of 0.586 across 5 evaluated AI models.
What OpenAI-MRCR: 2 needle 1M measures
OpenAI-MRCR: 2 needle 1M is a text benchmark that evaluates large language models on long context and reasoning tasks. LLM Stats tracks 5 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.4, with the leader reaching 0.6.
Compare leaders on the best AI for long context and best AI for reasoning leaderboards.
Publication
- Paper
- Michelangelo: Long Context Evaluations Beyond Haystacks via Latent Structure Queries
- Authors
- Kiran Vodrahalli, Santiago Ontanon, Nilesh Tripuraneni, Kelvin Xu, and 20 others
- Published
- arXiv
- 2409.12640
Abstract
We introduce Michelangelo: a minimal, synthetic, and unleaked long-context reasoning evaluation for large language models which is also easy to automatically score. This evaluation is derived via a novel, unifying framework for evaluations over arbitrarily long contexts which measure the model's ability to do more than retrieve a single piece of information from its context. The central idea of the Latent Structure Queries framework (LSQ) is to construct tasks which require a model to ``chisel away'' the irrelevant information in the context, revealing a latent structure in the context. To verify a model's understanding of this latent structure, we query the model for details of the structure. Using LSQ, we produce three diagnostic long-context evaluations across code and natural-language domains intended to provide a stronger signal of long-context language model capabilities. We perform evaluations on several state-of-the-art models and demonstrate both that a) the proposed evaluations are high-signal and b) that there is significant room for improvement in synthesizing long-context information.
MiniMax M1 40K leads with 58.6%, followed by MiniMax M1 80K at 56.2% and 
GPT-4.1 at 46.3%.
Progress Over Time
Interactive timeline showing model performance evolution on OpenAI-MRCR: 2 needle 1M
OpenAI-MRCR: 2 needle 1M Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  MiniMax | 456B | — | — | ||
| 2 | MiniMax | 456B | — | — | ||
| 3 | OpenAI | — | 1.0M | $2.00 / $8.00 | ||
| 4 | OpenAI | — | 1.0M | $0.40 / $1.60 | ||
| 5 | OpenAI | — | 1.0M | $0.10 / $0.40 |
FAQ
Common questions about OpenAI-MRCR: 2 needle 1M.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions