Qwen 2.5 Coder 32B

Overview

Alibaba's Qwen 2.5 Coder 32B — the strongest open-weight code model we've benchmarked, trading punches with Claude 3.5 Sonnet on HumanEval.

When to pick this model

• Self-hosted code copilots replacing proprietary APIs
• Repo-scale analysis and refactoring up to 128k tokens
• Polyglot codebases spanning dozens of languages
• Commercial code tooling needing Apache 2.0
• Generating production code where quality justifies the VRAM

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	19 GB
Q5_K_M	23 GB
Q8_0	35 GB
FP16 (no quantization)	64 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Published benchmark scores

Benchmark	Score
HumanEval	92.7
MBPP	86
LiveCodeBench	31.4

Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.

Strengths

• Best-in-class open-weight code generation
• Claude 3.5 Sonnet-level HumanEval scores
• 128k context for repo-wide tasks
• Apache 2.0 license

Limitations

• Requires 20+ GB VRAM at Q4
• Weaker than Qwen 2.5 32B for general chat
• Slower than 7B-class models for autocomplete loops

Architecture & training

Architecture: Dense Transformer specialized for code · 64 layers

Training: General pre-training + 5.5T code tokens, 92 languages.

Quick start

ollama run qwen2.5-coder:32b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.