Open Portuguese LLM - TweetSentBR: leaderboard
Sentiment-classification task from the Open Portuguese LLM Leaderboard (TweetSentBR), reporting macro-F1 over positive, negative, and neutral Brazilian Portuguese tweets.
Metric: F1 (%). Source: huggingface.co. Saturation forecast: Around October 2027. 42 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 79.28 | #123 |
| 2 | GPT-5 (Minimal) | 78.78 | #98 (GPT-5) |
| 3 | Claude 3.7 Sonnet (20250219) | 78.42 | #201 |
| 4 | Llama 3.1 405B Instruct | 78.21 | #457 |
| 5 | Gemini 2.5 Flash | 78.01 | #245 |
| 6 | GPT-4o (2024-08-06) | 77.61 | #336 |
| 7 | Gemini 2.5 Flash Lite | 76.96 | #422 |
| 8 | Gemini 2.0 Flash Lite (001) | 76.76 | #544 |
| 9 | Gemini 2.0 Flash (001) | 76.53 | #360 |
| 10 | Qwen 2.5 72B Instruct | 76.2 | #454 |
| 11 | DeepSeek V3.1 | 75.84 | #264 |
| 12 | GPT-4o Mini (2024-07-18) | 75.09 | #558 |
| 13 | Gemma 3 27B (IT) | 74.49 | #523 |
| 14 | Kimi K2 | 74.28 | #243 |
| 15 | GPT-5 Nano | 73.86 | #425 |
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-tweetsentbr · How It Works · Data refreshed daily, snapshot 2026-10-20.