ArmBench-LLM - Text Processing: leaderboard
ArmBench-LLM category: Armenian text processing, the mean accuracy on its Punctuation and Space Fix tasks (restoring punctuation and correcting word spacing). 0-100, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Not forecast. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 Pro | 87.91 | #70 |
| 2 | Gemini 2.5 Flash | 85.63 | #255 |
| 3 | Gemma 4 31B (IT) | 83.69 | #178 |
| 4 | Gemini 3 Flash (Preview) | 83.57 | #90 |
| 5 | GPT-5.4 Mini | 81.38 | #228 |
| 6 | Mistral Large 3 | 81.17 | #410 |
| 7 | Claude 3.7 Sonnet | 80.92 | #259 |
| 8 | Grok 4 Fast | 77.88 | #266 |
| 9 | Claude Sonnet 4 | 76.77 | #207 |
| 10 | Grok 4.1 Fast | 76.61 | #227 |
| 11 | Mistral Small 3.2 | 76.5 | #608 |
| 12 | GPT-5.4 Nano | 76.31 | #359 |
| 13 | Gemma 3 27B (IT) | 76.17 | #528 |
| 14 | Gemini 3 Pro (Preview) | 76.02 | #71 |
| 15 | Qwen 3.5 27B | 73.2 | #192 |
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-text-processing · How It Works · Data refreshed daily, snapshot 2026-11-05.