ArmBench-LLM v2 - Reading: leaderboard
ArmBench-LLM 5-category edition Reading score: mean across Belebele, Syndarin, Scientific, DREAM multiple-choice accuracy and SQuAD and MS MARCO (BLEU/100) in Armenian. 0-100, higher is better.
Metric: Mean of accuracy and BLEU tasks (0-100). Source: huggingface.co. Saturation forecast: Around August 2027. 32 models tracked. Updated 22 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Ayb-0.5-9B | 77.42 | |
| 2 | Ayb-0.5-4B | 73.75 | |
| 3 | GPT-6.1 Sol | 72.5 | #15 |
| 4 | DeepSeek V4.1 Flash | 71.18 | #69 |
| 5 | Ayb-0.5-0.8B | 69.33 | |
| 6 | Qwen 3.5 27B | 69.17 | #380 |
| 7 | Gemini 2.5 Pro | 68.79 | #280 |
| 8 | GPT-5.4 Mini | 68.31 | #409 |
| 9 | GPT-5.2 Pro | 67.84 | #139 |
| 10 | Claude Sonnet 4 | 67.71 | #414 |
| 11 | Gemma 4 31B (IT) | 67.58 | #354 |
| 12 | Claude 3.7 Sonnet | 67.53 | #455 |
| 13 | Gemini 3 Pro (Preview) | 67.36 | #123 |
| 14 | Grok 4 Fast | 67.05 | #446 |
| 15 | GLM-4.7 Flash | 66.4 | #1349 |
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-reading · How It Works · Data refreshed daily, snapshot 2026-11-22.