Update README.md

README.md

@@ -1,6 +1,10 @@
 ---
 library_name: transformers
-tags: []
+license: apache-2.0
+language:
+- en
+base_model:
+- Qwen/Qwen3-4B-Thinking-2507
 ---
 
 # QED-Nano
@@ -21,16 +25,10 @@ tags: []
 
 QED-Nano is a 4B parameter language model designed to push the theorem-proving boundaries of small models. It was post-trained to solve Olympiad-level proof problems by continually improving at test-time. When we scale test-time compute for our model, we get QED-Nano (Agent) which outperforms several much larger open source models (notably GPT-OSS-120B, Qwen3-235B-Thinking), and comes close to matching Gemini-3-Pro model on the challenging IMOProofBench benchmark:
 
-[ADD FINAL PLOT]
+![imoproofbench.png](https://huggingface.co/lm-provers/QED-Nano/resolve/main/imoproofbench.png)
 
 QED-Nano is based on Qwen/Qwen3-4B-Thinking-2507, and was post-trained via a combination of supervised fine-tuning and reinforcement learning with reasoning cache on a mixture of Olympiads proof problems from various sources. 
 
-### Key features
-- Reasoning model optimized for **theorem proving**
-- **Fully open model**: open weights + full training details including public data mixture and training configs
-
-For more details refer to our blog post: https://huggingface.co/spaces/lm-provers/qed-nano-blogpost
-
 ## How to use
 
 ```python