BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases8

AIME 2025 (AA)

As of 29 Oct 2026, the top score on AIME 2025 (AA) is 94.3% by Nova 2.0 Lite (Amazon), ahead of gpt-oss-120b at 93.4%, across 78 model configurations with a published result (Artificial Analysis, result dated 29 Oct 2025). All 30 problems of the 2025 American Invitational Mathematics Examination, olympiad-level with integer answers. Run by Artificial Analysis. AIME 2025 (AA) is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 29 Oct 2026)

  • Nova 2.0 Lite holds the top AIME 2025 (AA) score at 94.3% (Artificial Analysis, 29 Oct 2025).
  • gpt-oss-120b is second on AIME 2025 (AA) at 93.4% (Artificial Analysis, 5 Aug 2025).
  • AIME 2025 (AA): All 30 problems of the 2025 American Invitational Mathematics Examination, olympiad-level with integer answers. Run by Artificial Analysis (Artificial Analysis).
  • 78 model configurations have a published AIME 2025 (AA) result on BenchLeader (BenchLeader, 29 Oct 2026).
  • AIME 2025 (AA) does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

All 30 problems of the 2025 American Invitational Mathematics Examination, olympiad-level with integer answers. Run by Artificial Analysis.

Published by
Artificial Analysis
Category
Maths
Index weight
Reference only
Models
78
Data as of
29 Oct 2026
Newest result
4 Jan 2026

The publisher has posted no new results on this board since 4 Jan 2026. BenchLeader re-reads it every day, so newer models appear here as soon as they are published.

Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.

Top 15
  1. 1Nova 2.0 Lite (thinking)94.3%
  2. 2gpt-oss-120b (high)93.4%
  3. 3Nemotron 3 Nano 30B A3B (thinking)91.0%
  4. 4Nova 2.0 Omni (medium)89.7%
  5. 5gpt-oss-20b (high)89.3%
  6. 6Nova 2.0 Pro Preview (medium)89.0%
  7. 7Nova 2.0 Lite (medium)88.7%
  8. 8o388.3%
  9. 9Intellect 388.0%
  10. 10Apriel-v1.6-15B-Thinker88.0%
  11. 11ERNIE 5.0 (thinking)85.0%
  12. 12Qwen3-Next 80B A3B (thinking)84.3%
  13. 13Ring Flash 2.083.7%
  14. 14Magistral Medium 1.282.0%
  15. 15Magistral Small 1.280.3%
78 of 78
#
1Nova 2.0 Litethinking94.3%–2025-10-29
2gpt-oss-120bhigh93.4%49.82025-08-05
3Nemotron 3 Nano 30B A3Bthinking91.0%47.52025-12-15
4Nova 2.0 Omnimedium89.7%–2025-11-26
5gpt-oss-20bhigh89.3%45.62025-08-05
6Nova 2.0 Pro Previewmedium89.0%53.62025-11-27
7Nova 2.0 Litemedium88.7%–2025-10-29
8o388.3%61.02025-04-16
9Intellect 388.0%48.62025-11-27
10Apriel-v1.6-15B-Thinker88.0%–2025-11-25
11ERNIE 5.0thinking85.0%–2025-11-13
12Qwen3-Next 80B A3Bthinking84.3%52.32025-09-11
13Ring Flash 2.083.7%46.92025-09-19
14Magistral Medium 1.282.0%43.82025-09-18
15Magistral Small 1.280.3%40.02025-09-17
16Motif-2-12.7Bthinking80.3%–2025-12-04
17Falcon-H1R-7B80.0%–2026-01-04
18EXAONE 4.0 32Bthinking80.0%–2025-07-15
19Doubao-Seed-Code79.3%–2025-11-11
20K2 Think V2high78.3%–2025-12-05
21Olmo 3.1 32Bthinking77.3%45.22025-12-12
22Mi Dm K 2.5 Pro Dec2876.7%–2025-12-11
23Nemotron Nano 12B v2 VLthinking75.0%–2025-10-28
24Qwen3 Omni 30B A3B Instructthinking74.0%–2025-09-22
25Cogito v2.1thinking72.7%–2025-11-18
26Olmo 3 7B Thinkthinking70.7%–2025-11-20
27Hermes 4 - Llama-3.1 405Bthinking69.7%–2025-08-27
28NVIDIA Nemotron Nano 9B v2thinking69.7%–2025-08-18
29Hermes 4 70Bthinking68.7%–2025-08-27
30gpt-oss-120blow66.7%45.72025-08-05
31Qwen3-Next 80B A3B66.3%51.32025-09-11
32K2 Think V2medium64.7%–2025-12-05
33Llama Nemotron Ultra 253Bthinking63.7%–2025-04-07
34Nova 2.0 Pro Previewlow63.3%52.02025-11-27
35gpt-oss-20blow62.3%45.72025-08-05
36NVIDIA Nemotron Nano 9B v2no reasoning62.3%–2025-08-18
37HyperCLOVA X SEED Think (32B)thinking59.0%–2025-12-26
38Nova 2.0 Omnilow56.0%–2025-11-26
39Qwen3 Omni 30B A3B Instruct52.3%–2025-09-22
40Exaone 4.0 1.2Bthinking50.3%–2025-07-15
41Nova 2.0 Litelow46.7%–2025-10-29
42ERNIE 4.541.3%–2025-06-30
43Olmo 3 7B Think41.3%–2025-11-20
44EXAONE 4.0 32Bno reasoning39.3%–2025-07-15
45Nova 2.0 Omnino reasoning37.0%–2025-11-26
46Kimi Linear 48B A3B Instruct36.3%–2025-10-30
47K2 Think V2low35.3%–2025-12-05
48Nova 2.0 Liteno reasoning33.7%–2025-10-29
49Reka Flash 333.7%–2025-03-10
50Ministral 3 8B31.7%38.72025-12-02
51Nova 2.0 Pro Previewno reasoning30.7%–2025-11-27
52Ministral 3 14B30.0%39.92025-12-02
53Nemotron Nano 12B v2 VLno reasoning26.7%–2025-10-28
54Exaone 4.0 1.2Bno reasoning24.0%–2025-07-15
55Ministral 3 3B22.0%36.82025-12-02
56Llama 4 Maverick19.3%42.22025-04-05
57Phi-418.0%39.52024-12-12
58Nova Premier17.3%46.12025-04-30
59Hermes 4 - Llama-3.1 405Bno reasoning15.3%–2025-08-27
60Llama 4 Scout14.0%38.62025-04-05

More maths benchmarks · All LLM benchmarks

Related benchmarks

Cite as: BenchLeader, “AIME 2025 (AA) leaderboard”, data as of 29 Oct 2026, https://www.benchleader.com/benchmarks/aa_aime_2025.

AIME 2025 (AA): questions

What does AIME 2025 (AA) measure?
All 30 problems of the 2025 American Invitational Mathematics Examination, olympiad-level with integer answers. Run by Artificial Analysis. Scores are reported in percent of tasks solved; higher is better.
What is the top score on AIME 2025 (AA)?
As of 29 Oct 2026, the highest score on AIME 2025 (AA) is 94.3% by Nova 2.0 Lite (result dated 29 Oct 2025), ahead of gpt-oss-120b at 93.4%. Scores come from Artificial Analysis and are re-read daily.
Which AI model leads AIME 2025 (AA)?
Nova 2.0 Lite leads AIME 2025 (AA) with 94.3% as of 29 Oct 2026, ahead of gpt-oss-120b at 93.4% (result dated 29 Oct 2025).
How many models have AIME 2025 (AA) results?
78 model configurations have a AIME 2025 (AA) result on BenchLeader, all taken from Artificial Analysis.
Who runs AIME 2025 (AA) and how often is it updated?
AIME 2025 (AA) is published by Artificial Analysis. BenchLeader re-reads the published results once a day and records the date each result was published.
Does AIME 2025 (AA) count toward the BenchLeader Index?
No. AIME 2025 (AA) is shown for reference but left out of the composite index.