Evals for Every Language - ARC: leaderboard
LanguageBench ARC 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. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 98.79 | #58 |
| 2 | Claude Opus 4.8 | 98 | #37 |
| 3 | GPT-5.5 | 97.82 | #31 |
| 4 | DeepSeek V3.1 | 97.4 | #264 |
| 5 | Claude Opus 4.7 | 97.23 | #49 |
| 6 | Gemini 3.1 Flash Lite | 96.94 | #191 |
| 7 | GPT-5.4 | 96.36 | #79 |
| 8 | Qwen 3.6 Plus | 95.76 | #120 |
| 9 | Grok 4.20 | 95.65 | #175 |
| 10 | Claude Sonnet 4.6 | 95.55 | #90 |
| 11 | GPT-5.2 | 95.14 | #110 |
| 12 | GPT-5.3 | 95.06 | #103 |
| 13 | GPT-5 | 93.39 | #98 |
| 14 | GPT-5.1 | 92.43 | #135 |
| 15 | Claude Sonnet 4.5 | 92.24 | #142 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-arc · How It Works · Data refreshed daily, snapshot 2026-10-20.