Add known limitations (vision status)

README.md

@@ -35,6 +35,10 @@ AWQ 4-bit quantization of [Gemma 4 26B-A4B-it](https://huggingface.co/google/gem
 
 Standard community GPTQ under-calibrates rare experts due to routing imbalance. This model uses forced-routing calibration to ensure all 128 experts are properly quantized.
 
+## Known Limitations
+
+- **Vision: BROKEN** — Vision encoder layers (`embed_vision.*`) were quantized to INT4. This likely degrades or destroys vision quality. **Use for text-only inference.** A future version should add vision layers to `modules_to_not_convert`.
+
 ## Usage with SGLang
 
 ```bash