HELM AFR - Winogrande (Amharic): leaderboard
The "Winogrande (Amharic)" 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 | Claude 3.7 Sonnet (20250219) | 71.08 | #203 |
| 2 | Gemini 2.0 Flash | 69.27 | #344 |
| 3 | Gemini 2.0 Flash Lite | 62.71 | #453 |
| 4 | DeepSeek V3 | 61.63 | #323 |
| 5 | Llama 3.1 405B Instruct | 59.99 | #461 |
| 6 | GPT-4o (2024-08-06) | 59.42 | #335 |
| 7 | Claude 3.5 Haiku (20241022) | 54.39 | #591 |
| 8 | Llama 3.1 70B Instruct | 52.97 | #566 |
| 9 | Gemma 2 27B (IT) | 51.95 | #649 |
| 10 | Mixtral 8x7B Instruct | 51.73 | #1172 |
| 11 | Gemma 2 9B (IT) | 51.67 | #795 |
| 12 | GPT-3.5 Turbo (0125) | 51.1 | #988 |
| 13 | Llama 3.1 8B Instruct | 49.63 | #1040 |
| 14 | GPT-2 XL | 49.52 | #1617 |
| 15 | Mistral 7B Instruct (v0.3) | 49.46 | #1276 |
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-amharic · How It Works · Data refreshed daily, snapshot 2026-10-23.