IMO 2025
International Mathematical Olympiad 2025, proof-graded. MathArena.
As of 25 Oct 2026, GPT-5 leads IMO 2025 on BenchLeader with 38.1%, ahead of Gemini 2.5 Pro at 31.6%, across 6 model configurations with a published result.
MathArena marks 1 of these 6 results as possibly contaminated: the model was evaluated after the questions were public, so they may already be in its training data. Those scores are in the table below, and count in the index at the same weight as the rest.
- Published by
- MathArena
- Category
- Maths
- Index weight
- 0.5
- Models
- 6
- Data as of
- 25 Oct 2026
MIT — eth-sri/matharena.
What the test looks like
The six problems of the 2025 International Mathematical Olympiad, graded on full written proofs by MathArena's graders under the competition's marking scheme.
How it is scored
Percent of available marks, judged by human graders.
What to keep in mind
Six problems and human grading: a single problem is a sixth of the score. Different labs' own announcements used different conditions and are not comparable to this run.
- 1GPT-5 (high)38.1%
- 2Gemini 2.5 Pro31.6%
- 3Grok 421.4%
- 4o3 (high)16.7%
- 5o4-mini (high)14.3%
- 6DeepSeek R1 05286.8%
| # | ||||
|---|---|---|---|---|
| 1 |  GPT-5high | 38.1% | 57.9 | possible |
| 2 |  Gemini 2.5 Pro | 31.6% | 53.9 | – |
| 3 |  Grok 4 | 21.4% | 58.0 | – |
| 4 | o3high | 16.7% | 53.9 | – |
| 5 | o4-minihigh | 14.3% | 50.1 | – |
| 6 |  DeepSeek R1 0528open ↗ | 6.8% | 49.5 | – |
More maths benchmarks · All LLM benchmarks
Cite as: BenchLeader, “IMO 2025 leaderboard”, data as of 25 Oct 2026, https://www.benchleader.com/benchmarks/matharena_imo_2025.
IMO 2025: questions
- What does IMO 2025 measure?
- The six problems of the 2025 International Mathematical Olympiad, graded on full written proofs by MathArena's graders under the competition's marking scheme. Scores are reported in percent of tasks solved; higher is better.
- Which AI model leads IMO 2025?
- GPT-5 leads IMO 2025 with 38.1% as of 25 Oct 2026, ahead of Gemini 2.5 Pro at 31.6%.
- How many models have IMO 2025 results?
- 6 model configurations have a IMO 2025 result on BenchLeader, all taken from MathArena.
- Who runs IMO 2025 and how often is it updated?
- IMO 2025 is published by MathArena. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does IMO 2025 count toward the BenchLeader Index?
- Yes. IMO 2025 contributes to the maths category of the BenchLeader Index, normalised so that 50 is the average of the evaluated models.