Qwen2.5-Coder 32B Instruct

Permissive

Alibaba · Qwen · Released 2024-11

32BText

Widely regarded as the strongest fully open code model at its size, often cited as close to GPT-4-class coding performance.

Strengths

  • +Apache 2.0 license
  • +State-of-the-art open coding benchmark scores at release
  • +128K context for large-file/repo tasks

Limitations

  • -Needs a 24GB+ GPU for full-quality local inference
  • -Code-focused tuning trades off some general chat quality

License

Apache 2.0

use commercially with attribution niceties

Hardware

wants 24-48GB of VRAM - a 3090/4090-class card or better

Links

Stats

Qwen/Qwen2.5-Coder-32B-Instruct— downloads·— likes

via Hugging Face

Related models