SAHARA - Sentiment Analysis: leaderboard
Sentiment analysis of texts in African languages, from the SAHARA African-language LLM leaderboard (UBC-NLP, ACL 2025). The task score is the mean over the languages the task covers; Macro-F1 on a 0-100 scale, higher is better.
Metric: Macro-F1 (%). Source: huggingface.co. Saturation forecast: Around 2029. 48 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 57.52 | #258 |
| 2 | Gemini 3 Pro (Preview) | 55.68 | #71 |
| 3 | Claude Sonnet 4 (20250514) | 55.67 | #223 |
| 4 | Gemini 2.5 Pro | 53.01 | #159 |
| 5 | Gemini 2.5 Flash | 52.07 | #255 |
| 6 | GPT-5.1 (2025-11-13) | 51.56 | #107 |
| 7 | GPT-5 | 51.02 | #105 |
| 8 | GPT-5.6 Luna | 48.66 | #79 |
| 9 | Qwen 3.5 27B | 47.91 | #192 |
| 10 | c4ai-command-a-03-2025 | 47.14 | #468 |
| 11 | Qwen 3.6 27B | 46.58 | #182 |
| 12 | Gemma 3 27B (IT) | 46.35 | #528 |
| 13 | Llama 3.3 70B Instruct | 45.82 | #541 |
| 14 | Llama 4 Scout Instruct | 45.12 | #547 |
| 15 | Qwen 3.6 35B A3B | 44.68 | #238 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sahara-sentiment-analysis · How It Works · Data refreshed daily, snapshot 2026-11-05.