add

README.md

@@ -14,7 +14,7 @@ short_description: Model arena for software engineering tasks
 
 # SWE-Model-Arena
 
-An open-source platform for evaluating agentic coding models head-to-head. Both sides share the **same scaffolding** ([opencode](https://opencode.ai)) with identical tools, prompts, and sandboxed environments — the **only variable** is the underlying tool-calling model.
+An open-source platform for evaluating tool-calling models head-to-head. Both sides share the **same scaffolding** ([opencode](https://opencode.ai)) with identical tools, prompts, and sandboxed environments — the **only variable** is the underlying tool-calling model.
 
 **[Try it on Hugging Face Spaces](https://huggingface.co/spaces/SWE-Arena/SWE-Model-Arena)**
 

app.py

@@ -2053,7 +2053,7 @@ with gr.Blocks(title="SWE-Model-Arena", theme=gr.themes.Soft()) as app:
 
     with gr.Tab("⚔️Arena"):
         gr.Markdown("# ⚔️ SWE-Model-Arena")
-        gr.Markdown("Blind head-to-head agentic coding model comparison — same scaffold (opencode), different tool-calling LLM")
+        gr.Markdown("Blind head-to-head tool-calling model comparison — same scaffold (opencode), different tool-calling LLM")
 
         gr.Markdown("### 📜 How It Works")
         gr.Markdown(