Qwen 2.5 Coder
Partner VerifiedPublisher: Alibaba Cloud · Category: Coding
Parameters
32B
Context Window
128K
Default Size
20.2 GB
Downloads
412,800
Specialized code generation model trained on multi-trillion tokens of source code, documentation, and unit tests.
PULL MODEL
eq pull qwen2.5-coder:32b
RUN INFERENCE
eq run qwen2.5-coder:32b "Write a Rust async HTTP client"
Attention Architecture
Utilizes Grouped-Query Attention (GQA) with 16-token fixed Paged KV Cache block allocation for zero-fragmentation memory residency.
ONNX & EQC Compatible
Compiles directly through the EQC toolchain into standalone .eqx binary packages with fused SwiGLU kernels.
Multi-Hardware Support
Auto-detects CUDA RTX/A100/H100, Apple Silicon Metal Performance Shaders, or AVX-512 CPU execution backends.