AfroBench — leaderboard

AfroBench evaluates model capability on multilingual tasks from the linked upstream source with Average score as the primary reported metric.

Metric: Average Score (%). Source: huggingface.co. Status: saturation imminent. 24 models tracked.

Top models

#ModelScore
1GPT-577.74
2Gemini 3 Pro76
3Gemini 2.5 Pro74.53
4Claude Sonnet 4.570.6
5Claude Sonnet 468.7
6Gemini 2.5 Flash66.71
7Gemini 2.0 Flash66.43
8GPT-4o65.8
9GPT-4o (2024-08-06)65.8
10GPT-4.165.7
11Gemini 1.5 Pro62.8
12Claude 3.7 Sonnet60.3
13Gemma 3 27B51.01
14Gemma 2 27B44.9
15Gemma 2 9B40.8

Interactive version: theaggregate.ai/benchmark?slug=afrobench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.