You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+17Lines changed: 17 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -133,6 +133,7 @@
133
133
134
134
## 📑 Table of Contents
135
135
136
+
-[📰 News](#-news)
136
137
-[🚀 Key Features](#-key-features)
137
138
-[🏗️ Architecture](#️-architecture)
138
139
-[🚀 Quick Start](#-quick-start)
@@ -141,6 +142,22 @@
141
142
-[⭐ Star History](#-star-history)
142
143
-[📄 License](#-license)
143
144
145
+
146
+
---
147
+
148
+
## 📰 News
149
+
150
+
🎉 **[2025-10-28] DeepCode Achieves State-of-the-Art Performance on PaperBench Code-Dev!**
151
+
152
+
- 🏆 **Surpasses Human Experts**: DeepCode achieves **75.9%** on the 3-paper subset, outperforming **Top ML PhD** (72.4%) by **+3.5%**
153
+
- 🥇 **Outperforms Commercial Agents**: **+26.1%** improvement over best commercial code agents (**Cursor, Claude Code, Codex**) with **84.8%** accuracy
154
+
- 🔬 **Advances Scientific Code Generation**: **+22.4%** improvement over PaperCoder, the previous SOTA scientific code agent
155
+
- 🚀 **Beats LLM-Based Agents**: **+30.2%** improvement over best LLM agent frameworks, demonstrating the power of sophisticated agent architecture
0 commit comments