Overview

DeepCoder-14B-Preview was fine-tuned from DeepSeek-R1-Distill-Qwen-14B with distributed RL on verified coding problems; dataset, code and logs were released.

Highlights

  • 14B parameters (plus 1.5B version)
  • RL-trained on verifiable coding tasks
  • Fully open recipe
  • MIT license

Typical uses

Local competitive-programming style coding and research.

Things to consider

  • Preview release; long reasoning traces.
  • Download sizes refer to the default quantized versions in the Ollama library; full-precision weights are larger.

Specifications are taken from the model card on Hugging Face, the Ollama library and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.