Artificial Analysis AIME 2025 (AA AIME 2025)
We show this table for reference; we do not rank on it.
An independently evaluated AIME 2025 result from Artificial Analysis.
Benchmark score on AA AIME 2025 — October 29, 2026
We compile the AA AIME 2025 rows from secondary reports. GPT-5.2 leads the table at 99.0%, followed by Gemini 3 Pro (95.7%) and GLM-4.7 (95.0%). We do not use these results to rank models overall.
GPT-5.2
OpenAI
Gemini 3 Pro
GLM-4.7
Z.AI
13 modelsMathematicsCurrentDisplay onlyUpdated October 29, 2026
Benchmark score table (13 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | GPT-5.2OpenAI | 99.0% | Not reported | Closed |
| 2 | Gemini 3 ProGoogle | 95.7% | Not reported | Closed |
| 3 | GLM-4.7Z.AI | 95.0% | Not reported | Open |
| 4 | GPT-5 (high)OpenAI | 94.3% | Not reported | Closed |
| 5 | GPT-5.1OpenAI | 94.0% | Not reported | Closed |
| 6 | GPT-OSS 120BOpenAI | 93.4% | Not reported | Open |
| 7 | Grok 4xAI | 92.7% | Not reported | Closed |
| 8 | Claude Opus 4.5 ThinkingAnthropic | 91.3% | Not reported | Closed |
| 9 | o4-miniOpenAI | 90.7% | Not reported | Closed |
| 10 | GPT-5 miniOpenAI | 90.7% | Not reported | Closed |
| 11 | K-ExaoneLG AI Research | 90.3% | Not reported | Closed |
| 12 | DeepSeek V3.1 (Reasoning)DeepSeek | 89.7% | Not reported | Open |
| 13 | Grok 4 Fast (Reasoning)xAI | 89.7% | Not reported | Closed |
Among the reported AA AIME 2025 rows, GPT-5.2 is first at 99.0%. The third row is 4.0 points behind. The broader top-10 range is 8.3 points, so many of the published results sit in a relatively narrow band.
13 models have been evaluated on AA AIME 2025. The benchmark falls in the Mathematics category. AA AIME 2025 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About AA AIME 2025
Year
2026
Tasks
30 AIME 2025 problems
Format
Accuracy
Difficulty
Olympiad mathematics
Stored separately from the core AIME 2025 lane because Artificial Analysis controls the evaluation configuration.
Freshness and provenance
Version
AA AIME 2025 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does AA AIME 2025 measure?
An independently evaluated AIME 2025 result from Artificial Analysis.
Which model scores highest on AA AIME 2025?
GPT-5.2 by OpenAI currently leads with a score of 99.0% on AA AIME 2025.
How many models are evaluated on AA AIME 2025?
13 AI models have published results on AA AIME 2025 in the BenchLM catalog.
Compare top models on AA AIME 2025
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.