UX4567/Khushi-Business-AI-4bit

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Oct 2, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Khushi-Business-AI-4bit is a 7-billion parameter, 4-bit quantized conversational AI model developed by Kartik Sharma (UX4567). Built on Mistral 7B using Unsloth, it is optimized for low VRAM environments (6GB+ VRAM) and specifically fine-tuned for Indian business use cases. This model excels at tasks like customer support, sales chat, and lead handling, offering bilingual support for Hindi and English (Hinglish) within a 4096-token context length.

Loading preview...

Khushi-Business-AI-4bit (7B)

Khushi-Business-AI-4bit is a 7-billion parameter conversational AI model developed by Kartik Sharma (UX4567), built upon the Mistral 7B architecture. It is specifically designed for Indian business use cases and is optimized for efficiency through 4-bit quantization using Unsloth, allowing it to run on systems with 6GB+ VRAM.

Key Capabilities

  • Low VRAM Footprint: Quantized to 4-bit (BF16 base), requiring approximately 4GB RAM.
  • Business-Tuned: Optimized for tasks such as customer support, sales chat, and lead handling.
  • Bilingual Support: Understands and responds in both Hindi and English (Hinglish).
  • Fast Fine-tuning: Leverages Unsloth for 2x faster fine-tuning and quantization.
  • Flexible Formats: Available in safetensors and GGUF formats for compatibility with tools like Ollama and LM Studio.

Good for

  • Businesses in India requiring AI for customer interactions.
  • Developers needing a conversational AI that supports Hinglish.
  • Deployments on hardware with limited VRAM (e.g., consumer GPUs).
  • Applications focused on sales, support, and lead management in a business context.

Finetunes
1 models