AIME leaderboard
5 models have a published score. Reflection Beam 1.0 (Reflection AI) leads with 97.7, and the spread to the lowest score, Gemma 4 26B A4B at 88.3, is 9.4 points. The highest-scoring model you can deploy on OpenLLM Buddy is Qwen3.6 27B, ranked 3 of 5 with 94.1. Sources report different settings for this test, such as test year, version, or reasoning effort. Each setting is listed in the source table, so compare close scores with care.
Best closed api
No published score
0 models scored
AIME scores
Pass@1 accuracy, usually averaged over several samples per problem. No partial credit.
What it measures
Each AIME year has 30 problems (AIME I and II) with integer answers from 0 to 999, covering algebra, geometry, number theory, and combinatorics. Labs report a specific year, which is shown next to each score.
How it is scored
Pass@1 accuracy, usually averaged over several samples per problem. No partial credit.
Why it matters
Fresh contest years limit training-data contamination and reward careful multi-step math.
Official benchmark pageMore reasoning tests
Sources
| Model | Test | Score | Setting | Source |
|---|---|---|---|---|
| Reflection Beam 1.0 | AIME | 97.7 | AIME 2026, BF16 | Hugging Face model card (Beam-1.0-501B-A23B) |
| gpt-oss-120b | AIME | 95.8 | AIME 2024, high effort, no tools | gpt-oss model card paper (arXiv 2508.10925), Table 3 |
| Qwen3.6 27B | AIME | 94.1 | AIME 2026 | Qwen3.6-27B model card (Hugging Face) |
| Gemma 4 31B | AIME | 89.2 | AIME 2026, no tools | Gemma 4 31B IT model card (Hugging Face) |
| Gemma 4 26B A4B | AIME | 88.3 | AIME 2026, no tools | Gemma 4 31B IT model card (Hugging Face), 26B A4B column |
Scores are published results from model cards, lab reports, and public leaderboards, each linked to its source. Settings such as thinking mode or test version differ between sources and are shown next to each score. OpenLLM Buddy catalog models are uncensored fine-tunes; their scores are those of the official base model. Data collected 2026-10-30. How the index works