HELM AFR - Winogrande (Afrikaans): leaderboard
The "Winogrande (Afrikaans)" subtask of HELM AFR. Stanford HELM's African-languages project (mmlu-winogrande-afr): MMLU subject exams and Winogrande coreference translated into sixteen African languages, scored against the same items in English so a model's loss from the language switch...
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-08-06) | 81.44 | #335 |
| 2 | GPT-4 Turbo | 81.38 | #406 |
| 3 | Claude 3.7 Sonnet (20250219) | 78.61 | #203 |
| 4 | DeepSeek V3 | 78.27 | #323 |
| 5 | Gemini 2.0 Flash | 77.31 | #344 |
| 6 | Llama 3.1 405B Instruct | 76.06 | #461 |
| 7 | Qwen 2.5 72B Instruct | 75.1 | #458 |
| 8 | Llama 3.1 70B Instruct | 73.57 | #566 |
| 9 | Gemini 2.0 Flash Lite | 71.19 | #453 |
| 10 | Gemma 2 27B (IT) | 66.72 | #649 |
| 11 | Claude 3.5 Haiku (20241022) | 62.54 | #591 |
| 12 | Llama 3.3 70B Instruct | 59.2 | #536 |
| 13 | GPT-3.5 Turbo (0125) | 59.03 | #988 |
| 14 | Mixtral 8x7B Instruct | 56.88 | #1172 |
| 15 | Gemma 2 9B (IT) | 55.12 | #795 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-afr-winogrande-afrikaans · How It Works · Data refreshed daily, snapshot 2026-10-23.