Update README.md

README.md

@@ -34,6 +34,11 @@ This repository hosts the **open-source training and evaluation pipeline** as we
 **Repo purpose:** host the open-source training/eval pipeline and release artifacts.
 
 
+## Performance Snapshot
+<img src="figures/model_comparison_chart.png" width="720" />
+
+
+*Figure 1. DMind-3-nano significantly outperforms both the untuned base model and a similarly sized general-purpose model (Qwen3-0.6B), especially in multi-turn success.*
 
 ## Highlights
 
@@ -156,14 +161,6 @@ User: "查一下 Solana 上的 SOL"
 Model: <start_function_call>call:SEARCH_TOKEN{symbol:"SOL",chain:"solana"}<end_function_call>
 ```
 
-## Performance Snapshot
-
-- Function recognition: ~98.8% on validated set
-- Argument extraction: ~97.4%
-- Protocol adherence: SEARCH_TOKEN 98.5%, EXECUTE_SWAP 97.3%
-- Multi-turn success: ~93.7%
-
-Scope: tokens/chains listed in **Model Overview**; outside that set may be lower.
 
 ## License & Governance