Update README.md

README.md

@@ -27,10 +27,6 @@ tags:
 
 ## 🏆 Benchmarks
 
-![Benchmarks Table](https://huggingface.co/SupraLabs/Supra-50M_BASE/resolve/main/benchmarks.png)
-
-Supra-50M outperforms much larger models — GPT-2 Small (124M), SmolLM-135M, and OpenELM-270M — on BLiMP and SciQ, while remaining closely competitive on PIQA, HellaSwag, and ARC-Easy. All of this at a fraction of the parameter count.
-
 ### Benchmark Table
 
 | Benchmark | Supra-50M *(ours)* | GPT-2 (124M) | SmolLM-135M | OpenELM-270M |