ArmBench-LLM v2 - Correction: leaderboard
ArmBench-LLM 5-category edition Correction score: mean accuracy across Armenian Punctuation restoration and Space Fix tasks. 0-100, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Not forecast. 32 models tracked. Updated 22 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Ayb-0.5-9B | 90.55 | |
| 2 | GPT-6.1 Sol | 88.8 | #15 |
| 3 | GPT-5.2 Pro | 87.91 | #139 |
| 4 | Ayb-0.5-0.8B | 87 | |
| 5 | DeepSeek V4.1 Flash | 86.1 | #69 |
| 6 | Gemini 2.5 Flash | 85.63 | #466 |
| 7 | Ayb-0.5-4B | 84.85 | |
| 8 | Gemma 4 31B (IT) | 83.69 | #354 |
| 9 | Gemini 3 Flash (Preview) | 83.57 | #156 |
| 10 | GPT-5.4 Mini | 81.38 | #409 |
| 11 | Mistral Large 3 | 81.17 | #835 |
| 12 | Claude 3.7 Sonnet | 80.92 | #455 |
| 13 | Grok 4 Fast | 77.88 | #446 |
| 14 | Claude Sonnet 4 | 76.77 | #414 |
| 15 | Grok 4.1 Fast | 76.61 | #431 |
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-correction · How It Works · Data refreshed daily, snapshot 2026-11-22.