Alpha
Ctrl K
Contact
Assess · DeepSeek-V3

Can you own it?

Ownership levelPartialnone·limited·partial·substantial·fullAnalytical input C ยท 64.4/100

This page is a projection of the one entry record, the Use & modify and Transparency factors that Assess covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.

Which domain expands which factor
  • AssessUse & modify + Transparency
  • ImplementData control + Reliability
  • UseReliability
  • SupportTransparency

Intended & out-of-scope use

The original DeepSeek-V3 (December 2024) is the 671B-total / 37B-active Mixture-of-Experts Base and Chat model, 128K context, that established the V3 line. It is intended as a general assistant (Chat) or a foundation for further training (Base). It is a separate entry from the later MIT-licensed V3 generations because its weights are governed by the custom "DeepSeek License Agreement v1.0", not MIT.

Out-of-scope: any use that violates the licence's field-of-use restrictions (see License); any use requiring topic-neutral factuality (documented political censorship); the Base variant in any customer-facing role (it is not safety-tuned); and any unguarded high-stakes deployment. For new work, prefer the MIT V3 generations (deepseek-v3-mit) unless you specifically need this checkpoint.

Known limitations, bias & failure modes

Beyond the class-typical hallucination, the distinctive limitations are topic censorship aligned with Chinese content rules, lighter safety tuning than Western frontier labs, no first-party guard model, and - unique to this entry versus the MIT generations - a use-restricted licence that constrains what you may deploy it for. The Base variant is a raw completion model with no safety tuning. Recorded factually: China-origin model, which some organisations restrict by policy.

Openness tier & components

Original V3 sits in the open_weights tier (dimension ceiling 3). The weights are openly downloadable (ungated) and the documentation - a detailed technical report - is a strength, but the licence is conditional (use-restricted), training data is closed, and training code partial. The restrictions bite on legal and ownership rather than dropping the openness tier, because download is ungated and the grant is irrevocable.

License terms & what you may do

This is the load-bearing difference from the MIT generations. The code repository is MIT, but the weights are governed by the "DeepSeek License Agreement, Version 1.0" (non-OSI). The grant is genuinely generous in structure - "perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable", commercial use allowed, and "DeepSeek claims no rights in the Output You generate" - but it is conditioned on RAIL-style use restrictions (Section 5 / Attachment A): no use violating applicable law, no military application, no harm to minors, no generating verifiably false content to harm others, no unauthorised PII distribution, no defamation/harassment, no discrimination, no automated decisions affecting legal rights. These flow down to derivatives. The patent licence terminates if you sue DeepSeek over the work. Because of the field-of-use restrictions, use-and-modify is moderate and ownership is partial - one step below the MIT V3 generations.

Supply-chain provenance

The canonical source is the verified deepseek-ai organisation on Hugging Face, distributing safetensors with checksums and no malicious-checkpoint incident (checklist ~5/8). Safetensors is data-only, so loading the canonical weights cannot execute code. Short of a higher score only for the absence of cryptographic signing. The caveat: third-party quantizations are separate artifacts whose trust equals their uploader, and any redistribution must carry the DeepSeek License restrictions (not MIT). Pin the revision, verify the checksum, prefer the canonical org.

Open weights vs the hosted service. This entry documents the open weights, run on your own infrastructure - distinct from DeepSeek's hosted app and API, which has faced data-privacy scrutiny and bans in several jurisdictions. Those issues are not inherited by locally-run weights but are relevant if you call DeepSeek's own API.

EU AI Act posture

Original V3 is a GPAI model. Unlike the MIT generations, its weight licence is not free-and-open-source (it carries field-of-use restrictions), so the Article 53 open-source exemption does not apply on the licence axis at all. The surviving obligations are also unmet: no Article 55 documentation, no copyright policy, no training-content summary, and the corpus is closed. The technical report cites ~2.788M H800 GPU-hours; at 671B the systemic-risk question is live but not grounded on a disclosed FLOP figure. For a downstream EU deployer this is a concrete compliance gap, and the grade holds at C.

Benchmarks & evaluation

On public-leaderboard evidence the original V3 was competitive among large open-weight models at release. OneHill has not re-run these benchmarks this cycle, so performance is capped at 4 and no specific figures are asserted as verified. Evaluate against your own task, particularly where topic-neutral factuality matters.

How this scores

The ownership factors this domain covers, drawn from the one entry record.

1

Use and modify freelyCan you run, modify and adapt it with no gate and no field-of-use trap?

Moderate

The weight grant is irrevocable, royalty-free and permits commercial use and modification, but it is a non-OSI custom licence with binding RAIL-style field-of-use restrictions (no military, no unlawful use, no harm to minors, no automated legal-rights decisions, etc.) that flow down to derivatives - so use-and-modify is moderate, not strong. This is the material difference from the MIT V3 generations.

How this scores (AOI sub-dimensions)
Openness3/5how much is released - weights, data, code, licence - and how freelyOpen-weights tier: the weights are openly downloadable (ungated) with an open technical report, but the licence is conditional (use-restricted), the training data closed, the training code partial, evaluation partial.
Legal2/5how permissive and clean the licence is for real commercial useThe weight grant is irrevocable and permits commercial use, but it is a non-OSI custom licence with RAIL-style field-of-use restrictions on top of the family EU gap: no Article 55 documentation, no copyright policy, no training-content summary, and at 671B the open-source exemption does not apply (the licence is not FOSS).
2

TransparencyDo you know what it is: weights, training, behaviour, and legible terms?

Moderate

Weights are inspectable and there is a detailed technical report, but the training data and code are closed, and the weights carry China-aligned topic filtering you cannot inspect - open_weights, so moderate.

How this scores (AOI sub-dimensions)
Provenance3/5how well we can trace and verify what went into the modelVerified deepseek-ai org on Hugging Face, safetensors distribution with checksums and a clear canonical source with no malicious-checkpoint incident (checklist ~5/8).
Governance3/5how accountable and well-documented the publisher isActive, named publisher (DeepSeek) with a verified org and a track record of technical reports, meeting the score-3 anchor.
What this means for adoptionYou partially own self-hosted original DeepSeek-V3: self-hosting keeps your data yours under an irrevocable grant (data-control strong, reliability strong), but the DeepSeek License Agreement imposes RAIL-style field-of-use restrictions that flow down to derivatives, holding use-and-modify to moderate - so ownership is partial, one step below the MIT V3 generations (deepseek-v3-mit), which reach substantial. For new work prefer the MIT generations unless you specifically need this checkpoint; where you do use it, honour the use restrictions, deploy behind your own guardrails, and treat EU high-stakes use as needing a self-assembled compliance package (with no open-source exemption on the licence axis).

Sources

The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.

Licenceread2026-08-03
DeepSeek-V3 LICENSE-CODE / LICENSE-MODEL, read: "This code repository is licensed under the MIT License.
Model cardread2026-08-03
DeepSeek-V3 model card on the verified deepseek-ai HF org: 671B total / 37B active MoE, 128K context, safetensors, Base + Chat variants; first-class serving (vLLM, SGLang, llama.cpp, Ollama) and an extensive community quant ecosystem.
Third-party analysisunverified2026-08-03
On independent public leaderboards the original DeepSeek-V3 was competitive among large open-weight models at release (coding, maths, analysis); not OneHill-reproduced this session.
Third-party analysisunverified2026-08-03
Independent analysis notes DeepSeek open-weight models apply China-aligned content filtering on politically sensitive topics, with lighter safety tuning than Western frontier labs and no companion guard model.
Third-party analysisunverified2026-08-03
The DeepSeek License Agreement imposes field-of-use restrictions (non-FOSS), and no public EU AI Act training-content summary, copyright policy, or GPAI documentation package is published for DeepSeek-V3; the training corpus is not released.