dheeyantra/dhee-nxtgen-qwen3-indic

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Jan 8, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

Dhee-NxtGen-Qwen3-Indic is a 4 billion parameter multilingual large language model developed by DheeYantra in collaboration with NxtGen Cloud Technologies Pvt. Ltd., built on the Qwen3-4B architecture. This model is specifically designed for assistant-style conversations, reasoning, and function-calling workflows across 14 Indian (Indic) languages. It is optimized for native-script generation, consistent multilingual behavior, and cross-lingual generalization, making it ideal for AI applications targeting Indic language users.

Loading preview...

Dhee-NxtGen-Qwen3-Indic: Multilingual LLM for Indic Languages

Dhee-NxtGen-Qwen3-Indic is a 4 billion parameter multilingual large language model developed by DheeYantra and NxtGen Cloud Technologies Pvt. Ltd. Based on the Qwen3-4B architecture, this model is uniquely designed to support assistant-style conversations, reasoning, and function-calling across 14 Indian (Indic) languages within a single, unified model. It excels at native-script generation and maintains consistent multilingual behavior.

Key Capabilities

  • Single Multilingual Model: Supports 14 Indic languages (Hindi, Bengali, Tamil, Telugu, Malayalam, Gujarati, Kannada, Marathi, Odia, Punjabi, Assamese, Maithili, Sanskrit, Sindhi) without requiring per-language checkpoints.
  • Fluent Native-Script Generation: Optimized for generating text directly in the native scripts of supported languages.
  • Assistant-Style & Reasoning: Designed for conversational AI, summarization, Q&A, and long-form content generation.
  • Function-Calling Compatible: Supports prompting styles compatible with function/tool calling for advanced interactions.
  • Hugging Face & vLLM Compatible: Fully integrates with Hugging Face Transformers and is ready for high-throughput inference using vLLM.

Intended Uses

  • Multilingual Indic chatbots and AI assistants.
  • AI applications for education, governance, and the public sector in Indian languages.
  • Content generation and summarization in various Indian languages.
  • Cross-lingual conversational and reasoning systems.

Limitations

  • May occasionally produce inaccurate facts or hallucinate.
  • Performance can vary slightly across different languages.
  • Not suitable for medical, legal, or safety-critical applications.
  • Code-mixed inputs (e.g., Hinglish) may reduce output quality.

Finetunes
2 models