OpenCoder
A code LLM that ships its whole cookbook — data pipeline, ablations, and checkpoints included.
About
OpenCoder, from HPC-AI Tech, is an open code LLM whose headline feature is reproducibility: the team released not just weights but the full data-processing pipeline, the pretraining dataset recipe, and training protocols. The 8B instruct variant scores 83.5 on HumanEval(+) per the paper, on par with leading open code models. The HF README describes a permissive custom license allowing commercial use, though one third-party table lists MIT — confirm on the model card before shipping commercially.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Paper-reported: HumanEval(+) 83.5, MBPP(+) 79.1, BigCodeBench 40.3 (8B-Instruct).
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: Paper Nov 2024 (arXiv 2411.04905), HPC-AI Tech; Full reproducibility artifacts: data pipeline, protocols, checkpoints; License wording ambiguous between custom permissive and MIT. License: HF README describes a permissive custom license allowing commercial use; one third-party table lists MIT — confirm on the HF model card..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.0/10
- Maintenance · recency, releases, issue load
- 6.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Paper Nov 2024 (arXiv 2411.04905), HPC-AI Tech
- · Full reproducibility artifacts: data pipeline, protocols, checkpoints
- · License wording ambiguous between custom permissive and MIT