Open Portuguese LLM - ENEM: leaderboard
Multiple-choice questions from ENEM, Brazil's national upper-secondary exam (ENEM Challenge set). From the final table of the archived Open Portuguese LLM Leaderboard (3-shot, lm-evaluation-harness), covering about 1,210 open checkpoints and 42 API models; accuracy (%), higher is better. Frozen since the last evaluation on 2025-09-01.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 1253 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 97.69 | #126 |
| 2 | GPT-5 Mini (2025-08-07) | 95.66 | #175 |
| 3 | Gemini 2.5 Flash | 90.97 | #255 |
| 4 | Qwen 3 14B | 90.34 | #542 |
| 5 | GPT-5 Nano | 90.13 | #439 |
| 6 | Claude 3.7 Sonnet (20250219) | 89.01 | #209 |
| 7 | DeepSeek V3 (0324) | 89.01 | #349 |
| 8 | DeepSeek V3.1 | 88.87 | #275 |
| 9 | Kimi K2 | 87.89 | #252 |
| 10 | Gemini 2.0 Flash (001) | 87.89 | #365 |
| 11 | Llama 4 Maverick Instruct | 87.75 | #467 |
| 12 | Qwen 3 8B | 87.47 | #678 |
| 13 | Qwen 3 30B A3B | 87.4 | #496 |
| 14 | Qwen 2.5 VL 72B Instruct | 86 | #383 |
| 15 | Qwen 2.5 VL 32B Instruct | 86 | #463 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-portuguese-llm-enem · How It Works · Data refreshed daily, snapshot 2026-11-05.