AI BriefingAPK Reduced to Around 200MB
Key point
Pocket LLM v1.4.0 supports a smaller APK and in-app model downloads.
Details
Pocket LLM for Android v1.4.0 has been released, reducing the base APK to about 200MB.
Models are no longer bundled in the APK; instead, users select a model on first app launch and receive it via on-device download. Multiple downloaded models can be stored and switched between within the app.
Supported models are as follows.
- Gemma 4 E4B LiteRT
- Gemma 4 E2B LiteRT
- Qwen3 0.6B LiteRT
- Qwen3 0.6B Q4F16 ONNX
- Qwen2.5 0.5B ONNX
Additionally, the thinking text of supported models can now be collapsed, and the default UI has also been partially improved. All features are premised on fully offline use.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Toss Doctor V2 Released: Maximizing QA Efficiency with LLM-based Automated Testing and Self-healing