Evals for Every Language - TruthfulQA: leaderboard
LanguageBench TruthfulQA accuracy averaged across available languages and scaled to 0-100. Translation uses spBLEU only, without mixing chrF; higher is better.
Metric: Average Score (%). Source: huggingface.co. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 0 | #343 |
| 2 | Gemini 2.5 Pro | 0 | #152 |
| 3 | GPT-5 | 0 | #98 |
| 4 | Gemini 2.5 Flash | 0 | #245 |
| 5 | Claude Sonnet 4.5 | 0 | #142 |
| 6 | Llama 3.3 70B Instruct | 0 | #534 |
| 7 | GPT-4.1 | 0 | #248 |
| 8 | Llama 3.1 70B Instruct | 0 | #561 |
| 9 | GPT-OSS-120B | 0 | #341 |
| 10 | Claude Haiku 4.5 | 0 | #280 |
| 11 | GPT-5 Nano | 0 | #425 |
| 12 | Gemini 3 Pro (Preview) | 0 | #66 |
| 13 | GPT-5 Mini | 0 | #186 |
| 14 | Claude Sonnet 4 | 0 | #199 |
| 15 | Llama 4 Maverick | 0 | #464 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-truthfulqa · How It Works · Data refreshed daily, snapshot 2026-10-20.