MARCA (Portuguese, Basic): leaderboard
MARCA (2026) tests LLM web search on 52 manually authored multi-entity questions asked in Portuguese, in the Basic framework (the model itself calls web_search via the Serper API and web_scrape). A GPT-4.1 judge checks each answer against a manually validated checklist; checklist accuracy (%) averaged over three runs, higher is better.
Metric: Checklist accuracy (0-1, times 100): share of a question's checklist items (expected entities and attributes) that a GPT-4.1 judge marks satisfied, averaged over MARCA's 52 Portuguese questions and three runs, Basic framework (the model itself calls web_search via the Serper API and web_scrape); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 14 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 89.6 | #146 |
| 2 | GPT-5.2 | 85.5 | #190 |
| 3 | GPT-5 Mini | 85.2 | #373 |
| 4 | GPT-4.1 | 82.6 | #506 |
| 5 | Kimi K2 | 79.7 | #448 |
| 6 | Gemini 3 Flash | 73.3 | #180 |
| 7 | Sabiazinho-4 | 70.5 | #672 |
| 8 | Sabia-4 | 69.5 | #445 |
| 9 | Qwen 3 235B A22B | 67.7 | #633 |
| 10 | GPT-4.1 Mini | 63.1 | #799 |
| 11 | Gemini 2.5 Pro | 58.5 | #280 |
| 12 | Qwen 3 30B A3B | 48.4 | #1165 |
| 13 | Gemini 2.5 Flash | 47.5 | #466 |
| 14 | Gemini 2.5 Flash Lite | 29.9 | #972 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=marca-portuguese-basic · How It Works · Data refreshed daily, snapshot 2026-11-22.