Update src/about.py

src/about.py

@@ -86,6 +86,17 @@ This benchmark not only fills a critical gap in Persian LLM evaluation but also
  | Fairness   | 17%         |
  | Saftey     | 8.6%        |
  | Social norm| 74.4%       |
+
+## About Our Models
+
+In addition to the public models shown on this leaderboard, our company has developed 4 proprietary LLMs that are currently not publicly available:
+
+- **Ava**: Specialized in Persian RAG and normal tasks
+- **Gita**: Optimized for normal tasks with good reasoning
+- **Ahoran**: Our largest model with 8B param
+- **Sialk**: Lightweight model offering for normal Persian generations tasks
+
+These models remain unpublished due to [reason - e.g., internal deployment strategy, ongoing refinement].
 """
 
 EVALUATION_QUEUE_TEXT = """