LaboAI/LaboAI-0.3.2-1.5B

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 1, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

LaboAI-0.3.2-1.5B is a 1.5 billion parameter causal language model developed by LaboAI, fine-tuned specifically for Kotlin code generation, understanding, and debugging within Android development, with a strong focus on Jetpack Compose. Optimized using 4-bit QLoRA, it is designed for memory efficiency, enabling local execution on GPUs with limited VRAM. Its primary use case is as a private, low-latency coding assistant for Android developers, excelling at tasks like boilerplate generation, UI component creation, and debugging.

Loading preview...

LaboAI-0.3.2-1.5B: Specialized Kotlin & Android Code Assistant

LaboAI-0.3.2-1.5B is a compact 1.5 billion parameter language model developed by LaboAI, specifically fine-tuned for Kotlin code generation and Android development. Built upon the Qwen2.5-1.5B-Instruct base model, it emphasizes modern Android architectures, particularly Jetpack Compose.

Key Capabilities

  • Kotlin Code Generation: Generates boilerplate for Android components (Activities, Fragments, ViewModels).
  • Jetpack Compose: Creates modern UI components using Jetpack Compose.
  • Debugging: Assists in debugging compilation errors and logic flaws in Android code.
  • Java to Kotlin Translation: Translates legacy Java logic into modern Kotlin.
  • Memory Efficient: Optimized with 4-bit QLoRA for local execution on GPUs with limited VRAM (e.g., 4GB).

Good For

  • Local Coding Assistance: Ideal for use with tools like Ollama and the Continue extension in VS Code for private, low-latency code generation.
  • Android Developers: Specifically tailored for developers working with Kotlin and modern Android frameworks.
  • Privacy-Focused Workflows: Ensures code never leaves the local machine when used with local inference tools.

Limitations

  • Not optimized for general chat, creative writing, or complex mathematical reasoning.
  • May occasionally suggest deprecated Android APIs.
  • Has a limited context window of 2048 tokens, making it unsuitable for very large codebases.
  • Does not have real-time knowledge of the absolute latest Android library updates.

Finetunes
1 models