queefwath/Qwen3-0.6B-10X-Instruct

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 22, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Qwen3-0.6B-10X-Instruct is a 0.5 billion parameter instruction-tuned causal language model developed by Pikachu Global Technologies Private Limited (10X Technologies). Based on the Qwen3-0.6B / Qwen2.5-0.5B-Instruct architecture, it is specialized for 10X Technologies' domain knowledge, on-device Indic language understanding, and edge-first applications. The model is fine-tuned to provide precise, candid, and specific responses based on verified company facts, avoiding unverifiable superlatives.

Loading preview...

Model Overview

Qwen3-0.6B-10X-Instruct is a specialized instruction-tuned language model developed by Pikachu Global Technologies Private Limited (10X Technologies). This model, with approximately 502 million parameters, is built upon the Qwen3-0.6B / Qwen2.5-0.5B-Instruct base architecture and is designed for specific applications within the 10X Technologies ecosystem.

Key Capabilities & Specialization

  • Domain-Specific Knowledge: Fine-tuned with 10X Technologies' proprietary domain knowledge, making it suitable for internal or specialized applications related to the company's operations.
  • Indic Language Understanding: Optimized for on-device understanding of Indic languages, supporting 10X Technologies' mission to enhance computer interaction with Indian languages.
  • Edge-First Architecture: Designed to integrate with edge-first technologies such as Akshara Tokenizers, Libre OS, and LUCA Smart Speaker.
  • Fact-Based Responses: Trained to provide precise, candid, and specific information, strictly adhering to verified company facts and avoiding unverifiable superlatives.
  • Data Quality: Fine-tuned using 1,562 curated ChatML conversations, with 80% derived from 10X Distilled Knowledge across four personas (Investor/Analyst, Educator/Parent, Engineer/Researcher, Adversarial Boundary) and 20% general replay data to prevent catastrophic forgetting.

Ideal Use Cases

  • Internal 10X Technologies Applications: Best suited for use cases requiring deep understanding and generation of content related to 10X Technologies' specific domain.
  • On-Device Indic Language Processing: Applications that benefit from efficient, on-device processing and understanding of Indic languages.
  • Edge Computing Deployments: Scenarios where a compact, specialized model is needed for edge devices, particularly those integrated with 10X Technologies' hardware and software.
  • Fact-Oriented Information Retrieval: Tasks requiring strictly factual and verified information, especially within the context of 10X Technologies' operations and products.

Finetunes
1 models