- Cleus
- AI benchmarks
- AIME (OTIS mock exams)
AIME (OTIS mock exams)explained
Top high-school competition maths. The AIME is the exam the best maths students in the US sit to qualify for the national olympiad. Every answer is a whole number from 0 to 999, so there is no partial credit and no guessing. Epoch AI uses 45 questions from the 2024–2025 OTIS mock AIME exams, which are less likely to be in training data than the real exams.
- Highest published score
- Kimi K3.1 Pro, 100%
- Leader among popular models
- GPT-6 Astra
- Models with a result
- 210
- Skill it tests
- Math
What AIME (OTIS mock exams) really tests
- What it is
- The AIME is the exam the best maths students in the US sit to qualify for the national olympiad. Every answer is a whole number from 0 to 999, so there is no partial credit and no guessing. Epoch AI uses 45 questions from the 2024–2025 OTIS mock AIME exams, which are less likely to be in training data than the real exams.
- A question like
- How many ordered pairs of positive whole numbers (a, b) with a, b ≤ 100 make a² + b² divisible by 7? The answer is one number. Our own example in the style of the test, not a real item.
- Why it matters
- It is a clean test of multi-step reasoning: one slip anywhere and the final number is wrong.
- Reading the score
- Share of the 45 problems answered with the exact right number.
- The catch
- The best models now get nearly everything right, so it mainly separates weaker and older models. A model scoring 100% here can still fail at research maths.
- Made by
- Questions from OTIS mock AIME exams, run by Epoch AI. Official page
Leaderboard
Each model’s best published result across its reasoning settings.
Popular models
The models we track across this site.
Top 20 of all 210 models tested
Includes older, preview and specialist models that aren’t on our main leaderboard.
Questions
What is AIME (OTIS mock exams)?
The AIME is the exam the best maths students in the US sit to qualify for the national olympiad. Every answer is a whole number from 0 to 999, so there is no partial credit and no guessing. Epoch AI uses 45 questions from the 2024–2025 OTIS mock AIME exams, which are less likely to be in training data than the real exams.
What does a AIME (OTIS mock exams) score mean?
Share of the 45 problems answered with the exact right number.
Which AI model is best on AIME (OTIS mock exams)?
Kimi K3.1 Pro from Moonshot has the highest published AIME (OTIS mock exams) result we have (100%). Among the popular models on this site, GPT-6 Astra leads with 100%.
Is AIME (OTIS mock exams) a good benchmark?
It is a clean test of multi-step reasoning: one slip anywhere and the final number is wrong. The catch: The best models now get nearly everything right, so it mainly separates weaker and older models. A model scoring 100% here can still fail at research maths.
Who made AIME (OTIS mock exams)?
Questions from OTIS mock AIME exams, run by Epoch AI.
Benchmark scores: Epoch AI, Capabilities & benchmarking, used under CC BY 4.0. For each model we show its best published result across reasoning settings. Prices and context sizes: OpenRouter. Data refreshed November 9, 2026. Hands-on tests (answers, websites, code, refusals, speed) are our own single runs. Explanations written by Cleus.