Qwen3-0.6B — ONNX (INT4, onnxruntime-genai)

ONNX export of Qwen/Qwen3-0.6B in onnxruntime-genai format for efficient autoregressive text generation on CPU (INT4 RTN block-32 quantization).

Produced for use with inference4j, an inference-only AI library for Java.

Usage with inference4j

try (var generator = TextGenerator.builder()
        .model(ModelSources.qwen3_0_6B())
        .build()) {
    GenerationResult result = generator.generate("What is the capital of France?");
    System.out.println(result.text());
}

Model Details

Property	Value
Architecture	Qwen3 (0.6B parameters)
Task	Text generation / chat
Precision	INT4
ONNX format	onnxruntime-genai
License	Apache-2.0