Neura-Tech-AI/Nexa-AI-4B-Instruct

TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 11, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Nexa-AI-4B-Instruct is a 4 billion parameter instruction-tuned large language model developed collaboratively by Neura Tech AI and Lumina AI, built upon Qwen/Qwen3-4B-Instruct-2507. This model excels in general conversation, instruction following, coding, mathematics, and logical reasoning, featuring a substantial 262,144 token context length. It demonstrates strong performance across various benchmarks, particularly in knowledge, reasoning, and agent capabilities, and supports multilingual understanding.

Loading preview...

Nexa-AI-4B-Instruct Overview

Nexa-AI-4B-Instruct is a 4 billion parameter instruction-tuned large language model, a collaborative effort by Neura Tech AI and Lumina AI. It is based on the Qwen/Qwen3-4B-Instruct-2507 model and inherits its impressive 262,144 token context length, making it suitable for long-context understanding tasks.

Key Capabilities & Features

  • Multilingual AI Assistant: Strong performance in English, Hindi, Chinese, and other languages.
  • Instruction Following: High-quality responses to diverse instructions.
  • Coding Assistance: Capable of generating and understanding code.
  • Mathematical & Logical Reasoning: Excels in complex problem-solving.
  • Agent & Tool Calling: Supports advanced AI agent workflows and tool integration.
  • Fine-tuned Alignment: Designed for helpful and accurate responses.

Performance Highlights

Nexa-AI-4B-Instruct demonstrates competitive performance across several benchmarks, often outperforming its base model and other comparably sized models:

  • Knowledge: Achieves 69.6 on MMLU-Pro and 62.0 on GPQA, surpassing Qwen3-30B-A3B in some metrics.
  • Reasoning: Shows significant strength with 47.4 on AIME25 and 80.2 on ZebraLogic.
  • Coding: Scores 35.1 on LiveCodeBench v6 and 76.8 on MultiPL-E.
  • Agent: Leads with 61.9 on BFCL-v3 and strong results on TAU benchmarks.
  • Creative Writing: Achieves 83.5 on Creative Writing v3 and 83.4 on WritingBench.

Good For

  • Applications requiring robust instruction following and general conversational abilities.
  • Development of AI agents and systems that utilize tool calling.
  • Tasks involving coding, mathematical problem-solving, and logical reasoning.
  • Multilingual applications needing strong understanding across various languages.
  • Scenarios benefiting from long-context processing capabilities.

Finetunes
2 models