Skip to content

Brazilian AI

Where does Brazil stand in the AI race?

Updated · Rogue

Short answer

The United States and China hold the largest models and most of the compute. On public agent leaderboards, the score goes to the solved task, and there Rogue, a Brazilian company, appears in the same table as OpenAI, Google, Anthropic, Z.ai, Moonshot AI and DeepSeek. The Brazilian government plans up to R$ 23 billion for AI through 2028.

In the same table

Agent leaderboards score the solved task, wherever the lab is based. These are the numbers, with each lab's country.

Terminal-Bench 4.0
Model or systemLabCountryScore
Rogue ConstellationRogueBrazil
71.21%
Claude Opus 5.5AnthropicUnited States
66.4%
GPT-6 AstraOpenAIUnited States
57.9%
Gemini 4 ArgonGoogleUnited States
57.4%

Pass rate on Terminal-Bench 4.0 as published by each lab. Read the research

DeepSWE v1.1
Model or systemLabCountryScore
Rogue DeepThink PreviewRogueBrazil
75.22%
GPT-6 AstraOpenAIUnited States
74.12%
Gemini 3.8 FlashGoogleUnited States
73.83%
Claude Opus 5AnthropicUnited States
73.65%
GLM 5.3Z.aiChina
68.96%
Kimi K3Moonshot AIChina
68.51%
Grok 4.6xAIUnited States
66.74%
DeepSeek V4 ProDeepSeekChina
62.83%
Qwen 3.8 MaxAlibabaChina
57.46%

Best result per lab on the public DeepSWE v1.1 leaderboard (Datacurve, September 22, 2026), and Rogue DeepThink Preview. Read the research

What puts a country in the table

Training a frontier model takes compute that few countries have today. Agents, evaluation and systems built on those models take something else: careful research, evaluation data and engineering. That is the layer where Rogue works, with results that enter the same table as the frontier labs.

Brazil is investing in the foundation as well. The Brazilian Artificial Intelligence Plan provides up to R$ 23 billion through 2028, a supercomputer at LNCC in Petrópolis and Portuguese language models. SoberanIA, trained in Portuguese, launched under that plan in December 2025.

The map of AI in Brazil

Who does what, based on what each company publishes about itself.

CompanyTypeWhat it doesSource
RogueAgent research and productAgents that build websites, web apps, slides and games, plus Rogue Code for Mac. Research published with score, cost and conditions on Terminal-Bench and DeepSWE.Rogue research
Maritaca AIPortuguese language modelsThe Sabiá family (Sabiá 4, Sabiazinho 4, Sabiá 4 Thinking) and S5, specialized in Portuguese and Brazilian contexts, with a focus on legal work. Its technical report positions the models on cost-performance.maritaca.aiSabiá-4 report
WideLabsPortuguese language modelsMaker of Amazônia IA, a language model built in Portuguese with Brazilian cultural data, delivered as full-stack AI projects, mostly for the public sector.Mobile Time, Feb 2026
AdaptaAI hubPresents itself as “all the best AIs in one place”: ONE gives access to other companies' models, such as ChatGPT, Claude and Gemini, in one subscription with courses; Skip builds software without code. Its site lists no model of its own and no research.adapta.org
SoberanIAPublic projectA model trained in Portuguese, launched in December 2025 by the Ministry of Science, the Government of Piauí, Telebras and partners, as part of the Brazilian AI Plan.DPL News, Dec 2025

Sources checked on November 6, 2026.

Questions

Does Brazil have its own AI?

Yes. There are models trained in Portuguese, such as Sabiá, Amazônia IA and SoberanIA, and agent research published on global leaderboards, such as Rogue's.

How much is Brazil investing in AI?

The Brazilian Artificial Intelligence Plan provides up to R$ 23 billion through 2028.

Does a Brazilian company appear on global AI leaderboards?

Yes. Rogue publishes results on Terminal-Bench 4.0, Terminal-Bench-Science 0.1 and DeepSWE v1.1, next to models from OpenAI, Google, Anthropic, Z.ai, Moonshot AI, DeepSeek and Alibaba.

Sources

Related questions

See what the research builds

Open Rogue