Mathematics · Benchmark profile
American Invitational Mathematics Examination 2026 contest problems used as a frontier math evaluation.
Data verified 15 July 2026 · Methodology 1.2.0
Visual analysis
Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.
6 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| GLM-5.2 | #1 | Z.AI | 99.2% |
| GLM-5 | #2 | Z.AI | 95.8% |
| Kimi K2.5 | #3 | Moonshot AI | 95.8% |
| GLM-5.1 | #4 | Z.AI | 95.3% |
| Qwen3.6 Plus | #5 | Alibaba Cloud | 95.3% |
| Claude Opus 4.5 | #6 | Anthropic | 95.1% |
Shows how quickly scores separate down the field
| Model | Rank | Provider | Score |
|---|---|---|---|
| GLM-5.2 | #1 | Z.AI | 99.2% |
| GLM-5 | #2 | Z.AI | 95.8% |
| Kimi K2.5 | #3 | Moonshot AI | 95.8% |
| GLM-5.1 | #4 | Z.AI | 95.3% |
| Qwen3.6 Plus | #5 | Alibaba Cloud | 95.3% |
| Claude Opus 4.5 | #6 | Anthropic | 95.1% |
Average within the visible cohort · descriptive only
| Provider | Models | Average score |
|---|---|---|
| Z.AI | 3 | 96.8% |
| Moonshot AI | 1 | 95.8% |
| Alibaba Cloud | 1 | 95.3% |
| Anthropic | 1 | 95.1% |
One best score per model · higher is better
| Rank | Model | Provider | License | Score |
|---|---|---|---|---|
| #1 | GLM-5.2 glm-5-2 | Z.AI | open | 99.2% |
| #2 | GLM-5 glm-5 | Z.AI | open | 95.8% |
| #3 | Kimi K2.5 kimi-k2.5 | Moonshot AI | — | 95.8% |
| #4 | GLM-5.1 glm-5-1 | Z.AI | open | 95.3% |
| #5 | Qwen3.6 Plus qwen3-6-plus | Alibaba Cloud | closed | 95.3% |
| #6 | Claude Opus 4.5 claude-opus-4-5 | Anthropic | closed | 95.1% |
The top of this snapshot is led by GLM-5.2 at 99.2%; third place is 3.4 points behind. The top-6 spread is 4.1 points.
About AIME 2026
American Invitational Mathematics Examination 2026 contest problems used as a frontier math evaluation. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
American Invitational Mathematics Examination 2026 contest problems used as a frontier math evaluation.
GLM-5.2 by Z.AI currently leads with 99.2%.
6 models in the LuminaBench cohort have a qualifying score on this benchmark.
Yes. This family is ranking-weighted at 2.3% of its capability category. GLM-5.2 is overall #11.
Related
