# 0.9.0
## Release Information
- OS support - Tested with Windows-x86 (10), Windows-arm64 (10), Mac-arm64 (14), Linux (Ubuntu 22.04)
- QAIRT SDK - Tested with versions (2.42.0, 2.43.0)
- Model Visualization format support - ONNX, TensorFlow, PyTorch (1.x, 2.x), TFLite, Executorch Program, DLC, EAIX, HTP
## Highlights
- Model diff - CLI and Python API support
- Execute model diffing via a new command line interface, or through the Python API. View the diff report from the CLI interface in a QAIRT Visualizer window, or export it to JSON and load it directly in the full application.
- HTP op source mapping - adds support to map ops and tensors between DLC and HTP models
- Side-by-side graph view of DLC and HTP models to visualize translation of operations after context binary generation
- Click of an op or tensor in one graph highlights the corresponding op(s) and tensor(s) in the other graph (if exists)
- If the model was converted with source mappings included, all three graphs are visualized side by side and linked
- Safetensor Support
- Safetensor visualization in the graph panel is now supported, with the tensor data represented as nodes in the graph. Safetensor files may also be diffed against other Safetensor files.
- UI enhancements
- DLC large model support: Visualizing most DLC models of over 10 GB is now supported
## Known Issues
- Model Diff
- Diffing larger models in the QAIRT Visualizer Graph Panel (i.e., with 40K nodes or more or larger than 1.5 GB) may result in performance degradation or may fail to load altogether due to memory constraints. Upcoming releases will address these limitations to allow extra-large models to be loaded. This limit does not affect diffing models via the Python API’s diff() module or qairt-visualizer diff.
Last Published: Jun 03, 2026
[Previous Topic
Known Issues](https://docs.qualcomm.com/bundle/publicresource/80-87189-1/topics/0_8_0.md) [Next Topic
0.10.0](https://docs.qualcomm.com/bundle/publicresource/80-87189-1/topics/0_10_0.md)