Update README.md

README.md

@@ -12,9 +12,9 @@ base_model:
 
 #### *Update on 26.05.2025*
 
-*For those interested in building NLP applications for Estonian: Llammas 🐑 is no longer the best open-weight model for Estonian. For example, models from the Llama 3.1 family can understand and generate Estonian much better. For more details on the performance of various open LLMs on Estonian, check out our recent [paper](https://aclanthology.org/2025.nodalida-1.37/).*
+*For those interested in building NLP applications for Estonian: Llammas 🐑 is no longer the best open-weight model available for it. For instance, Llama 3.1 8B performs better out of the box. For a more detailed comparison of open LLMs on Estonian, see our recent [paper](https://aclanthology.org/2025.nodalida-1.37/).*
 
-*For an up-to-date leaderboard of both proprietary and open models on Estonian, please visit [baromeeter.tartunlp.ai](https://baromeeter.tartunlp.ai) (and consider contributing if you speak Estonian ☺️)*
+*For an up-to-date leaderboard of both proprietary and open models, please visit [baromeeter.tartunlp.ai](https://baromeeter.tartunlp.ai) (and consider contributing if you speak Estonian ☺️)*
 
 # LLammas 🐑