Open Portuguese LLM - OAB Exams: leaderboard
Multiple-choice questions from the Brazilian Bar Association (OAB) law exams. 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: Around February 2027. 1253 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 92.16 | #126 |
| 2 | Gemini 2.5 Flash | 83.55 | #255 |
| 3 | Claude 3.7 Sonnet (20250219) | 83.55 | #209 |
| 4 | GPT-4o (2024-08-06) | 82 | #343 |
| 5 | GPT-5 (Minimal) | 81.05 | #105 (GPT-5) |
| 6 | Gemini 2.0 Flash (001) | 77.68 | #365 |
| 7 | Llama 3.1 405B Instruct | 76.4 | #470 |
| 8 | Llama 4 Maverick Instruct | 72.85 | #467 |
| 9 | GPT-5 Mini (2025-08-07) | 71.85 | #175 |
| 10 | DeepSeek V3 (0324) | 71.48 | #349 |
| 11 | Llama 3.2 90B Vision Instruct | 70.62 | #621 |
| 12 | Gemini 2.0 Flash Lite (001) | 70.62 | #552 |
| 13 | DeepSeek V3.1 | 70.39 | #275 |
| 14 | Kimi K2 | 69.7 | #252 |
| 15 | Llama 3.1 70B | 68.97 | #588 |
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-oab-exams · How It Works · Data refreshed daily, snapshot 2026-11-05.