ArmBench-LLM v2 - Knowledge: leaderboard
ArmBench-LLM 5-category edition Knowledge score: mean across Armenian Unified Exams (History, Literature, Math rescaled from 0-20), INCLUDE, Hartak MCQA, Simple QA (BLEU/100) and Armenian MMLU-Pro. 0-100, higher is better.
Metric: Mean of exam (0-20 rescaled), accuracy and BLEU tasks (0-100). Source: huggingface.co. Saturation forecast: Around February 2027. 32 models tracked. Updated 22 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6.1 Sol | 77.94 | #15 |
| 2 | DeepSeek V4.1 Flash | 70.94 | #69 |
| 3 | GPT-5.2 Pro | 69.52 | #139 |
| 4 | Gemini 3 Pro (Preview) | 64.71 | #123 |
| 5 | Gemini 3 Flash (Preview) | 63.91 | #156 |
| 6 | Grok 4 Fast | 63.05 | #446 |
| 7 | Claude 3.7 Sonnet | 59.72 | #455 |
| 8 | Grok 4.1 Fast | 59.08 | #431 |
| 9 | Qwen 3.5 27B | 58.22 | #380 |
| 10 | Gemini 3.1 Pro (Preview) | 56.09 | #102 |
| 11 | Claude Sonnet 4 | 54.21 | #414 |
| 12 | Gemma 4 31B (IT) | 54.17 | #354 |
| 13 | GPT-5.4 Mini | 53.47 | #409 |
| 14 | Gemini 2.5 Pro | 53.36 | #280 |
| 15 | Mistral Large 3 | 50.92 | #835 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=armbench-llm-v2-knowledge · How It Works · Data refreshed daily, snapshot 2026-11-22.