AIME 2025
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.
GPT-5.2 Pro from OpenAI currently leads the AIME 2025 leaderboard with a score of 1.000 across 113 evaluated AI models.
What AIME 2025 measures
AIME 2025 is a text benchmark that evaluates large language models on reasoning and math tasks. LLM Stats tracks 113 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.8, with the leader reaching 1.0.
Compare leaders on the best AI for reasoning and best AI for math leaderboards.
Publication
- Paper
- Challenging the Boundaries of Reasoning: An Olympiad-Level Math Benchmark for Large Language Models
- Authors
- Haoxiang Sun, Yingqian Min, Zhipeng Chen, Wayne Xin Zhao, and 1 others
- Published
- arXiv
- 2503.21380
Abstract
The rapid advancement of large reasoning models has saturated existing math benchmarks, underscoring the urgent need for more challenging evaluation frameworks. To address this, we introduce OlymMATH, a rigorously curated, Olympiad-level math benchmark comprising 350 problems, each with parallel English and Chinese versions. OlymMATH is the first benchmark to unify dual evaluation paradigms within a single suite: (1) natural language evaluation through OlymMATH-EASY and OlymMATH-HARD, comprising 200 computational problems with numerical answers for objective rule-based assessment, and (2) formal verification through OlymMATH-LEAN, offering 150 problems formalized in Lean 4 for rigorous process-level evaluation. All problems are manually sourced from printed publications to minimize data contamination, verified by experts, and span four core domains. Extensive experiments reveal the benchmark's significant challenge, and our analysis also uncovers consistent performance gaps between languages and identifies cases where models employ heuristic "guessing" rather than rigorous reasoning. To further support community research, we release 582k+ reasoning trajectories, a visualization tool, and expert solutions at https://github.com/RUCAIBox/OlymMATH.
GPT-5.2 Pro leads with 100.0%, followed by GPT-5.2 at 100.0% and 
Kimi K2-Thinking-0905 at 100.0%.
Progress Over Time
Interactive timeline showing model performance evolution on AIME 2025
AIME 2025 Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | OpenAI | — | — | — | ||
| 1 | OpenAI | — | 400K | $1.75 / $14.00 | ||
| 1 | Moonshot AI | 1.0T | — | — | ||
| 1 |  xAI | — | — | — | ||
| 1 |  Google | — | — | — | ||
| 6 |  Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 7 | Google | — | 1.0M | $0.50 / $3.00 | ||
| 8 | OpenAI | — | — | — | ||
| 8 |  Meituan | 560B | — | — | ||
| 10 | 32B | 262K | $0.06 / $0.24 | |||
| 11 | OpenAI | 21B | — | — | ||
| 12 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 13 |  ByteDance | — | — | — | ||
| 14 |  StepFun | 196B | 66K | $0.10 / $0.40 | ||
| 15 |  Microsoft | 1.0T | — | — | ||
| 16 | OpenAI | — | — | — | ||
| 16 |  Sarvam AI | 30B | — | — | ||
| 16 | Sarvam AI | 105B | — | — | ||
| 19 | Moonshot AI | 1.0T | — | — | ||
| 20 |  DeepSeek | 685B | — | — | ||
| 21 |  Zhipu AI | 358B | — | — | ||
| 22 | OpenAI | — | — | — | ||
| 22 | OpenAI | — | — | — | ||
| 24 |  Xiaomi | 309B | — | — | ||
| 25 | OpenAI | — | — | — | ||
| 25 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 25 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 28 | Zhipu AI | 357B | — | — | ||
| 29 | xAI | — | 128K | $3.00 / $15.00 | ||
| 30 | DeepSeek | 685B | — | — | ||
| 30 | DeepSeek | 685B | — | — | ||
| 32 | ByteDance | — | — | — | ||
| 33 |  LG AI Research | 236B | — | — | ||
| 34 | OpenAI | — | — | — | ||
| 35 | OpenAI | 117B | 131K | $0.10 / $0.50 | ||
| 36 |  Amazon | — | — | — | ||
| 36 |  Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 38 | Amazon | — | — | — | ||
| 39 | xAI | — | 2.0M | $0.20 / $0.50 | ||
| 40 | xAI | — | — | — | ||
| 41 | Zhipu AI | 30B | — | — | ||
| 42 | OpenAI | — | 400K | $0.25 / $2.00 | ||
| 42 |  Inception | — | 128K | $0.25 / $0.75 | ||
| 44 | Amazon | — | 1.0M | $0.30 / $2.50 | ||
| 45 | xAI | — | — | — | ||
| 46 | Meituan | 560B | — | — | ||
| 47 | 120B | — | — | |||
| 48 | Alibaba Cloud / Qwen Team | 236B | — | — | ||
| 49 | DeepSeek | 685B | — | — | ||
| 50 | OpenAI | — | — | — |
FAQ
Common questions about AIME 2025.
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.
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
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from programming contests (LeetCode, AtCoder, CodeForces) and evaluates four different scenarios: code generation, self-repair, code execution, and test output prediction. Problems are annotated with release dates to enable evaluation on unseen problems released after a model's training cutoff.