Tier 2 Restricted open-weight
What the publisher claims: open-source, open-weight
The one-liner: Same lab, different licence: the code line ships under DeepSeek custom terms rather than the MIT used for V3 and R1.
The six questions
weights downloadable, ungated
non-OSI custom licence with use-based restrictions
deployable with standard tooling
architecture, training stages, token counts and evals documented; corpus withheld
Moderate for research replication, not exact reproduction.
GitHub repo accepts issues but development direction and releases are fully controlled by DeepSeek.
License
DeepSeek License Agreement v1.0 (model licence), distinct from the MIT used by DeepSeek V3/R1 — license source ↗
- custom permissive-with-restrictions licence: redistribution allowed with the same use-based restrictions carried into derivatives, plus prohibited uses (illegal content, PII, defamation, discrimination, military use, fully automated decisions affecting legal rights)
What is actually released
- weights (DeepSeek-Coder 1.3B/5.7B/6.7B/33B, V2 Lite 16B and V2 236B MoE)
- tech reports
- code
- evaluation results; no training data or corpus
Why it matters
Materially distinct from the DeepSeek text-family entry in the dataset; shows that openness must be read per model, not per vendor.
Sources
- https://raw.githubusercontent.com/deepseek-ai/DeepSeek-Coder-V2/main/LICENSE-MODEL
- https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Instruct
Derived from the openness dataset behind But Is It Open? — snapshot 2026-11-07 ↗. Scores are qualitative readings of the cited evidence, not verdicts; tier measures openness, not intelligence.