math benchmark
AIME 2026 Leaderboard
AIME 2026 is a math benchmark comprising all 30 problems from the 2026 American Invitational Mathematics Examination (AIME I and AIME II), with integer answers from 000-999.
Updated Sep 5, 2026
Models22
Model coverage22
MetricScore
EvidenceB
AIME 2026 Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelZA | Score99.20% | Percentile100.00% | Participants22 | EvidenceC | Evaluated |
| Rank02 | ModelSA | Score96.67% | Percentile95.24% | Participants22 | EvidenceC | Evaluated |
| Rank03 | ModelMA | Score96.40% | Percentile90.48% | Participants22 | EvidenceC | Evaluated |
| Rank04 | ModelTM | Score95.50% | Percentile85.71% | Participants22 | EvidenceC | Evaluated |
| Rank05 | ModelZA | Score95.30% | Percentile80.95% | Participants22 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score95.30% | Percentile76.19% | Participants22 | EvidenceC | Evaluated |
| Rank07 | ModelUP | Score95.30% | Percentile71.43% | Participants22 | EvidenceC | Evaluated |
| Rank08 | ModelME | Score94.70% | Percentile66.67% | Participants22 | EvidenceC | Evaluated |
| Rank09 | ModelMI | Score94.50% | Percentile61.90% | Participants22 | EvidenceC | Evaluated |
| Rank10 | ModelBY | Score94.20% | Percentile57.14% | Participants22 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score94.10% | Percentile52.38% | Participants22 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score92.70% | Percentile47.62% | Participants22 | EvidenceC | Evaluated |
| Rank13 | ModelLA | Score92.60% | Percentile42.86% | Participants22 | EvidenceC | Evaluated |
| Rank14 | ModelMI | Score92.50% | Percentile38.10% | Participants22 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score91.30% | Percentile33.33% | Participants22 | EvidenceC | Evaluated |
| Rank16 | ModelGO | Score89.20% | Percentile28.57% | Participants22 | EvidenceC | Evaluated |
| Rank17 | ModelGO | Score88.30% | Percentile23.81% | Participants22 | EvidenceC | Evaluated |
| Rank18 | ModelBY | Score88.30% | Percentile19.05% | Participants22 | EvidenceC | Evaluated |
| Rank19 | ModelGO | Score77.50% | Percentile14.29% | Participants22 | EvidenceC | Evaluated |
| Rank20 | ModelGO | Score69.10% | Percentile9.52% | Participants22 | EvidenceC | Evaluated |
| Rank21 | ModelGO | Score42.50% | Percentile4.76% | Participants22 | EvidenceC | Evaluated |
| Rank22 | ModelGO | Score37.50% | Percentile0.00% | Participants22 | EvidenceC | Evaluated |
AIME 2026 Highlights
The leading models and scores on this benchmark.
AIME 2026 Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for AIME 2026
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is AIME 2026?
What AIME 2026 measures and how its scores work.
AIME 2026 is an AI benchmark based on the 30 problems from the 2026 American Invitational Mathematics Examination (AIME I and AIME II).
It measures performance on complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- AIME 2026
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about AIME 2026.
Which model scores highest on AIME 2026?
GLM-5.2 is currently ranked first with 99.20%.
What are the top three models on AIME 2026?
The current leaders are GLM-5.2 (99.20%), Sakana Namazu (96.67%), and Kimi K2.6 (96.40%).
Which AIME 2026 model has the lowest official input price?
Qwen3.6-35B-A3B has the lowest matched official input price at $0.25 input / $1.5 output per 1M tokens.
Which models are fastest among AIME 2026 results?
The fastest matched records are Gemma 4 31B (22.86 tok/s via FriendliAI), Seed 2.0 Pro (4.40 tok/s via DeepInfra), and GLM-5.2 (3.96 tok/s via ZAI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.