BrainboxAI/code-il-E4B-safetensors

VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Apr 18, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

BrainboxAI/code-il-E4B-safetensors is a 7.9 billion parameter coding assistant model, fine-tuned by BrainboxAI, based on the Gemma 4 architecture. This repository provides the full 16-bit safetensors weights of the code-il-E4B model, designed for advanced use cases like continued training and deployment conversions. It excels at code generation and can respond in Hebrew while keeping code in English, offering a 32768 token context length.

Loading preview...

Overview

BrainboxAI/code-il-E4B-safetensors is a companion repository to the main BrainboxAI/code-il-E4B model, providing the full 16-bit weights in safetensors format. Developed by Netanel Elyasi of BrainboxAI, this 7.9 billion parameter model is a fine-tune of unsloth/gemma-4-E4B-it and is specifically designed as a coding assistant. Unlike the compressed version in the main repository, this build is intended for developers working on the model rather than just running it.

Key Capabilities

  • Full Precision Weights: Contains the complete 16-bit weights, ideal for advanced model manipulation.
  • Code Generation: Optimized for generating code, as demonstrated by its ability to implement complex algorithms like binary search.
  • Multilingual Interaction: Can understand and respond to user prompts in Hebrew, while consistently generating code in English.
  • Continued Training: Serves as the recommended starting point for further training on private codebases, preserving existing model knowledge.

When to Use This Build

  • Loading the model with transformers in Python for custom applications.
  • Continuing training or fine-tuning the model on specific datasets.
  • Converting the model to other deployment formats like ONNX.
  • Integrating with frameworks that do not support GGUF formats.

Important Note

For general use, such as running the model with Ollama or LM Studio, the compressed version in the main repository (BrainboxAI/code-il-E4B) is recommended due to its smaller size and ease of use.

Finetunes
1 models