ArmBench-LLM - POS: leaderboard
ArmBench-LLM category: Armenian part-of-speech tagging accuracy, evaluated with Metric AI Lab's lighteval fork. 0-100, higher is better; several otherwise capable models score at or near 0.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 80 | #178 |
| 2 | Claude Sonnet 4 | 77 | #207 |
| 3 | GLM-5 | 77 | #149 |
| 4 | Gemini 2.5 Flash | 76 | #255 |
| 5 | Gemini 2.5 Pro | 75 | #159 |
| 6 | GPT-5.4 Mini | 75 | #228 |
| 7 | Gemini 3 Flash (Preview) | 75 | #90 |
| 8 | Gemini 3 Pro (Preview) | 74 | #71 |
| 9 | Qwen 3.5 27B | 74 | #192 |
| 10 | Gemini 3.1 Pro (Preview) | 73 | #64 |
| 11 | Claude 3.7 Sonnet | 72 | #259 |
| 12 | Grok 4.1 Fast | 72 | #227 |
| 13 | Grok 4 Fast | 71 | #266 |
| 14 | GPT-5.2 Pro | 70 | #70 |
| 15 | GLM-4.7 Flash | 67 | #543 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=armbench-llm-pos · How It Works · Data refreshed daily, snapshot 2026-11-05.