AfroBench-Lite
Models complete language understanding, question answering, reasoning, and generation tasks across 14 African languages. Seven datasets use task-specific metrics, including accuracy, F1, character-level translation scores, and exact match.
This reduced task and language selection differs from the full 64-language AfroBench. The release month refers to the opened June 2025 paper revision. Per-task metrics must remain separate.
- Made by
- Masakhane, McGill University, Mila, University of Waterloo, University College London, Lelapa AI, and The African Research Collective
- Released
- June 2025
- Kind
- Knowledge in many languages
- Level
- School
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Many languages
- Interaction
- Single turn
- Items from
- Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- AfroBench
- In views
- Many languages
- Strongest model tested
-
Gemini 3 Pro #56 on the capability index
- Last model added
- November 2025
- Scores refreshed
- November 18th, 2026
In the authors’ words
Large-scale multilingual evaluations, such as MEGA, often include only a handful of African languages due to the scarcity of high-quality evaluation data and the limited discoverability of existing African datasets. This lack of representation hinders comprehensive LLM evaluation across a diverse range of languages and tasks. To address these challenges, we introduce AfroBench – a multi-task benchmark for evaluating the performance of LLMs across 64 African languages, 15 tasks and 22 datasets. AfroBench consists of nine natural language understanding datasets, six text generation datasets, six knowledge and question answering tasks, and one mathematical reasoning task. We present results comparing the performance of prompting LLMs to fine-tuned baselines based on BERT and T5-style models. Our results suggest large gaps in performance between high-resource languages, such as English, and African languages across most tasks; but performance also varies based on the availability of monolingual data resources. Our findings confirm that performance on African languages continues to remain a hurdle for current LLMs, underscoring the need for additional efforts to close this gap. https://mcgill-nlp.github.io/AfroBench/
GPT-5 leads with 77.7 among models released by November 2025 GPT-5 leads with 77.7 among models released by November 2025
Best score per entry from the benchmark's own leaderboard, read on November 18th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in November 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 77.7 GPT-5
- 76.0
Gemini 3 Pro
- 74.5
Gemini 2.5 Pro
- 70.6 Claude Sonnet 4.5
- 68.7 Claude Sonnet 4
- 66.7
Gemini 2.5 Flash
- 66.4
Gemini 2.0 Flash
- 65.8 GPT-4o (Aug)
- 65.7 GPT-4.1
- 62.8
Gemini 1.5 Pro
- 60.3 Claude 3.7 Sonnet
- 51.0
Gemma 3 27B
- 49.7 LLaMa 4 405B
- 44.9
Gemma 2 27B
- 43.4 Aya-101 13B
- 40.8
Gemma 2 9B
- 40.0
Llama 3.1-70B
- 28.8 LLaMAX3 8B
- 25.8
Llama 3.1-8B
- 23.7 Gemma1.1 7b
- 21.4
Llama 3-8B
- 21.0 Lugha-Llama 8B
- 16.2
Llama 2-7B
- 16.1 AfroLLaMa 8B
- 1 GPT-5 77.7
- 2
Gemini 3 Pro 76.0
- 3
Gemini 2.5 Pro Jun 2025 74.5
- 4 Claude Sonnet 4.5 70.6
- 5 Claude Sonnet 4 68.7
- 6
Gemini 2.5 Flash Sep 2025 66.7
- 7
Gemini 2.0 Flash Feb 2025 66.4
- 8 GPT-4o (Aug) 65.8
- 9 GPT-4.1 65.7
- 10
Gemini 1.5 Pro Sep 2024 62.8
- 11 Claude 3.7 Sonnet 60.3
- 12
Gemma 3 27B 51.0
- 13 LLaMa 4 405B 49.7
- 14
Gemma 2 27B 44.9
- 15 Aya-101 13B 43.4
- 16
Gemma 2 9B 40.8
- 17
Llama 3.1-70B 40.0
- 18 LLaMAX3 8B 28.8
- 19
Llama 3.1-8B 25.8
- 20 Gemma1.1 7b 23.7
- 21
Llama 3-8B 21.4
- 22 Lugha-Llama 8B 21.0
- 23
Llama 2-7B 16.2
- 24 AfroLLaMa 8B 16.1
Source: Official AfroBench-Lite leaderboard
- 77.7 GPT-5
- 76.0
Gemini 3 Pro
- 74.5
Gemini 2.5 Pro
- 70.6 Claude Sonnet 4.5
- 68.7 Claude Sonnet 4
- 66.7
Gemini 2.5 Flash
- 66.4
Gemini 2.0 Flash
- 65.7 GPT-4.1
- 62.8
Gemini 1.5 Pro
- 60.3 Claude 3.7 Sonnet
- 51.0
Gemma 3 27B
- 44.9
Gemma 2 27B
- 40.8
Gemma 2 9B
- 40.0
Llama 3.1-70B
- 25.8
Llama 3.1-8B
- 21.4
Llama 3-8B
- 16.2
Llama 2-7B
- 1 GPT-5 77.7
- 2
Gemini 3 Pro 76.0
- 3
Gemini 2.5 Pro Jun 2025 74.5
- 4 Claude Sonnet 4.5 70.6
- 5 Claude Sonnet 4 68.7
- 6
Gemini 2.5 Flash Sep 2025 66.7
- 7
Gemini 2.0 Flash Feb 2025 66.4
- 8 GPT-4.1 65.7
- 9
Gemini 1.5 Pro Sep 2024 62.8
- 10 Claude 3.7 Sonnet 60.3
- 11
Gemma 3 27B 51.0
- 12
Gemma 2 27B 44.9
- 13
Gemma 2 9B 40.8
- 14
Llama 3.1-70B 40.0
- 15
Llama 3.1-8B 25.8
- 16
Llama 3-8B 21.4
- 17
Llama 2-7B 16.2
Source: Official AfroBench-Lite leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 77.7 | Aug 2025 | Official leaderboard | |
| 2 | | 76.0 | Nov 2025 | Official leaderboard | |
| 3 | | 74.5 | Jun 2025 | Official leaderboard | |
| 4 | Claude Sonnet 4.5 | 70.6 | Sep 2025 | Official leaderboard | |
| 5 | Claude Sonnet 4 | 68.7 | May 2025 | Official leaderboard | |
| 6 | | 66.7 | Sep 2025 | Official leaderboard | |
| 7 | | 66.4 | Feb 2025 | Official leaderboard | |
| 8 | GPT-4o (Aug) | 65.8 | Official leaderboard | ||
| 9 | GPT-4.1 | 65.7 | Apr 2025 | Official leaderboard | |
| 10 | | 62.8 | Sep 2024 | Official leaderboard | |
| 11 | Claude 3.7 Sonnet | 60.3 | Feb 2025 | Official leaderboard | |
| 12 | | 51.0 | Mar 2025 | Official leaderboard | |
| 13 | LLaMa 4 405B | 49.7 | Official leaderboard | ||
| 14 | | 44.9 | Jun 2024 | Official leaderboard | |
| 15 | Aya-101 13B | 43.4 | Official leaderboard | ||
| 16 | | 40.8 | Jun 2024 | Official leaderboard | |
| 17 | | 40.0 | Jul 2024 | Official leaderboard | |
| 18 | LLaMAX3 8B | 28.8 | Official leaderboard | ||
| 19 | | 25.8 | Jul 2024 | Official leaderboard | |
| 20 | Gemma1.1 7b | 23.7 | Official leaderboard | ||
| 21 | | 21.4 | Apr 2024 | Official leaderboard | |
| 22 | Lugha-Llama 8B | 21.0 | Official leaderboard | ||
| 23 | | 16.2 | Jul 2023 | Official leaderboard | |
| 24 | AfroLLaMa 8B | 16.1 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 77.7 | Aug 2025 | Official leaderboard | |
| 2 | | 76.0 | Nov 2025 | Official leaderboard | |
| 3 | | 74.5 | Jun 2025 | Official leaderboard | |
| 4 | Claude Sonnet 4.5 | 70.6 | Sep 2025 | Official leaderboard | |
| 5 | Claude Sonnet 4 | 68.7 | May 2025 | Official leaderboard | |
| 6 | | 66.7 | Sep 2025 | Official leaderboard | |
| 7 | | 66.4 | Feb 2025 | Official leaderboard | |
| 8 | GPT-4.1 | 65.7 | Apr 2025 | Official leaderboard | |
| 9 | | 62.8 | Sep 2024 | Official leaderboard | |
| 10 | Claude 3.7 Sonnet | 60.3 | Feb 2025 | Official leaderboard | |
| 11 | | 51.0 | Mar 2025 | Official leaderboard | |
| 12 | | 44.9 | Jun 2024 | Official leaderboard | |
| 13 | | 40.8 | Jun 2024 | Official leaderboard | |
| 14 | | 40.0 | Jul 2024 | Official leaderboard | |
| 15 | | 25.8 | Jul 2024 | Official leaderboard | |
| 16 | | 21.4 | Apr 2024 | Official leaderboard | |
| 17 | | 16.2 | Jul 2023 | Official leaderboard |
Rankings by part
We read 7 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Intent, MMLU and Math. GPT-5 leads 4 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Intent
17 models we track have a score here.
- 1
Gemini 3 Pro ↑1 88.8
- 2
Gemini 2.5 Pro Jun 2025 ↑1 88.0
- 3
Gemini 2.5 Flash Sep 2025 ↑3 87.4
- 4 GPT-5 ↓3 87.3
- 5 GPT-4.1 ↑3 82.9
- 6 Claude Sonnet 4.5 ↓2 79.3
- 7 Claude Sonnet 4 ↓2 77.9
- 8
Gemini 1.5 Pro Sep 2024 ↑1 76.3
- 9
Gemini 2.0 Flash Feb 2025 ↓2 74.4
- 10 Claude 3.7 Sonnet 72.2
MMLU
17 models we track have a score here.
- 1
Gemini 3 Pro ↑1 86.1
- 2 GPT-5 ↓1 83.3
- 3
Gemini 2.5 Pro Jun 2025 77.4
- 4 Claude Sonnet 4.5 74.0
- 5 Claude Sonnet 4 72.7
- 6
Gemini 2.5 Flash Sep 2025 67.3
- 7 Claude 3.7 Sonnet ↑3 65.3
- 8
Gemini 1.5 Pro Sep 2024 ↑1 62.6
- 9 GPT-4.1 ↓1 58.6
- 10
Gemini 2.0 Flash Feb 2025 ↓3 58.0
Math
17 models we track have a score here.
- 1 GPT-5 73.7
- 2
Gemini 3 Pro 73.4
- 3
Gemini 2.5 Pro Jun 2025 73.2
- 4 Claude Sonnet 4.5 69.7
- 5
Gemini 2.5 Flash Sep 2025 ↑1 69.3
- 6
Gemini 2.0 Flash Feb 2025 ↑1 66.5
- 7 Claude Sonnet 4 ↓2 63.3
- 8
Gemini 1.5 Pro Sep 2024 ↑1 57.7
- 9 GPT-4.1 ↓1 57.3
- 10
Gemma 3 27B ↑1 44.6
Topic
17 models we track have a score here.
- 1 GPT-5 88.5
- 2
Gemini 2.5 Pro Jun 2025 ↑1 87.9
- 3
Gemini 3 Pro ↓1 87.9
- 4
Gemini 2.5 Flash Sep 2025 ↑2 86.8
- 5
Gemini 2.0 Flash Feb 2025 ↑2 84.8
- 6 Claude 3.7 Sonnet ↑4 84.5
- 7 GPT-4.1 ↑1 84.3
- 8 Claude Sonnet 4.5 ↓4 83.3
- 9 Claude Sonnet 4 ↓4 82.4
- 10
Gemini 1.5 Pro Sep 2024 ↓1 81.3
Natural language inference
17 models we track have a score here.
- 1 GPT-5 83.3
- 2
Gemini 3 Pro 77.4
- 3
Gemini 2.5 Pro Jun 2025 72.5
- 4 Claude Sonnet 4.5 69.9
- 5
Gemini 2.5 Flash Sep 2025 ↑1 69.3
- 6 GPT-4.1 ↑2 66.6
- 7
Gemini 2.0 Flash Feb 2025 66.2
- 8 Claude Sonnet 4 ↓3 66.2
- 9
Gemini 1.5 Pro Sep 2024 63.6
- 10 Claude 3.7 Sonnet 58.8
Machine translation
Translation from English or French into the evaluated languages. 17 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑6 57.0
- 2
Gemini 3 Pro 47.3
- 3 GPT-4.1 ↑5 47.1
- 4
Gemini 2.5 Pro Jun 2025 ↓1 46.3
- 5
Gemini 2.5 Flash Sep 2025 ↑1 45.3
- 6 Claude Sonnet 4.5 ↓2 45.2
- 7 GPT-5 ↓6 44.8
- 8 Claude Sonnet 4 ↓3 44.6
- 9
Gemini 1.5 Pro Sep 2024 43.0
- 10 Claude 3.7 Sonnet 42.9
Reading comprehension
17 models we track have a score here.
- 1 GPT-5 83.3
- 2
Gemini 2.5 Pro Jun 2025 ↑1 76.4
- 3 Claude Sonnet 4 ↑2 73.7
- 4 Claude Sonnet 4.5 72.8
- 5
Gemini 3 Pro ↓3 71.2
- 6 Claude 3.7 Sonnet ↑4 64.2
- 7 GPT-4.1 ↑1 62.9
- 8
Gemma 3 27B ↑3 60.1
- 9
Gemini 2.0 Flash Feb 2025 ↓2 58.1
- 10
Gemini 1.5 Pro Sep 2024 ↓1 55.0
3 models have held number one on the official leaderboard since its release in June 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official AfroBench-Lite leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official AfroBench-Lite leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Multilingual
- MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
- WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET.
K-EXAONE 90.5%
- MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language.
Qwen3.7-Max 87%
- CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories.
Solar Open 2 90.7%
- Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
- MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 12 months ago, in November 2025, for Gemini 3 Pro. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (24 entries).
- Pass Models scored 17 models we track, plus 7 other entries on its leaderboard.
- Warning Newest score The newest came 12 months ago, in November 2025, for Gemini 3 Pro.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3 Pro (#56).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 17 rows for models we track and 7 other entries at the last read.
- Pass Last read November 18th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.