Qwen 2.5 Coder 14B Instruct

Overview

Alibaba's Qwen 2.5 Coder 14B under Apache 2.0 with HumanEval 89.6 and LiveCodeBench 37.1. The VRAM sweet spot for serious self-hosted code generation.

When to pick this model

• Self-hosted coding agents on a single 24GB GPU
• Repo-scale code generation needing 131k context
• Permissively licensed alternative to Codestral
• Multi-language production codebases

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	9 GB
Q5_K_M	11 GB
Q8_0	16 GB
FP16 (no quantization)	28 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	89.6
MBPP	86.2
LiveCodeBench	37.1

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

Strengths

• HumanEval 89.6 — competitive with much larger coders
• LiveCodeBench 37.1
• Apache 2.0 license
• 131k context for long-file work

Limitations

• Weaker than general 14B models on non-code chat
• No vision input
• Outscored by frontier closed APIs on the hardest benchmarks

Architecture & training

Architecture: Dense 14B code · FIM

Training: 5.5T tokens of code.

Quick start

ollama run qwen2.5-coder:14b

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