Update README.md

README.md

@@ -34,6 +34,8 @@ We introduce the updated version of the [**T-pro-it-2.0**](https://huggingface.c
 T-pro-it-2.1 — is an efficient russian model built upon the Qwen 3 model family with improved instruction following and tool-calling capabilities compared to [T-pro-it-2.0](https://huggingface.co/t-tech/T-pro-it-2.0). 
 Outperforms Qwen3-32B in tool calling scenarios, which is essential for agentic applications. Built for both general tasks and complex workflows.
 
+More train details in our Habr: https://habr.com/ru/companies/tbank/articles/979650/
+
 **NOTE: This model supports only non-thinking mode and does not generate `<think></think>` blocks in its output. Meanwhile, specifying `enable_thinking=False` is no longer required.**
 
 ### 📚 Dataset
@@ -81,6 +83,8 @@ Tool-calling Tasks: Performance on par with Qwen3-235B-2507 on tool-calling benc
 
 \* IFeval metric is mean of 4 values: prompt and instruct levels for strict and loose accuracy.
 
+More benchmarks can be found in our [Habr post](https://habr.com/ru/companies/tbank/articles/979650/).
+
 ## Recommended Generation Parameters
 
 ```