---
title: "DeepSeek V3 family | Ethical Stack"
url: https://ethicalstack.ai/models/deepseek-v3/
catalog_version: f6baff2119d0e9158520da7fbe4232011dac79f25bd08757436d7a1dbd593db2
---

# DeepSeek V3 family

General-purpose MoE base and chat releases that underpin the separate R1 reasoning family; the original V3 card documents 671B total and 37B active parameters.

Canonical profile: https://ethicalstack.ai/models/deepseek-v3/

Scope: DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Published overall: 4.2; 5/6 dimensions rated.

Round the equal-weight mean of available numerical ratings to one decimal; require at least four of six for an overall score. BDS status stays outside the mean; for eligible profiles, a boycott target caps the final score at 2/10 and a pressure target subtracts one point (minimum 1). Not listed and not checked do not adjust the score.

## Energy practices

Finding: 2/10

Assessment: https://ethicalstack.ai/models/deepseek-v3/#energy

This is a practices and disclosure score for the named release, not a request-level energy or emissions estimate.

Reason: 1 + 1 reporting + 0 electricity + 0 controls = 2/10. This grades documented practices and transparency.

Argument: The V3 report gives a training-compute estimate in H800 GPU-hours for its final training run; this is not a direct electricity, emissions, or water measurement. The rubric rewards documented reporting and controls; it does not infer high consumption from missing measurements.

Type: argument; confidence: low

Scope (dimension): DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: No comparable per-request electricity, emissions, water, regional serving mix, or independently reviewed model footprint is established in the cited sources.

```json
{
  "method": "practices-transparency-v1",
  "scope": "DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.",
  "components": {
    "disclosure": {
      "score": 1,
      "reason": "The V3 report gives a training-compute estimate in H800 GPU-hours for its final training run; this is not a direct electricity, emissions, or water measurement.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    },
    "electricity": {
      "score": 0,
      "reason": "No model-specific clean-electricity sourcing or hourly power evidence is published in the reviewed sources; zero means unestablished, not dirty electricity.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    },
    "demand": {
      "score": 0,
      "reason": "No model-specific documented control or measured physical resource reduction is established by the reviewed release materials. Public weights or smaller variants alone do not prove reduced resource use.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    }
  },
  "actualImpactMeasured": false
}
```

- [Official model card or model release repository](https://huggingface.co/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Official source repository](https://github.com/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Developer technical report](https://arxiv.org/abs/2412.19437) (primary; checked 2026-10-24)

## Training data

Finding: 3/10

Assessment: https://ethicalstack.ai/models/deepseek-v3/#training-rights

The release provides some training-process information, but not a complete permission ledger for inherited and later training data.

Reason: 1 base + 2 source transparency + 0 documented permission + 0 creator controls = 3/10. Weight licensing is not treated as corpus permission.

Argument: The technical report describes a broad data pipeline and training stages, but does not disclose a full source-level corpus and permission ledger for the inherited pretraining mixture. A public weight license governs use of those weights; it is not proof that every training source was permitted.

Type: evidence; confidence: low

Scope (dimension): DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Complete source lists, source-specific rights, inherited-base provenance, synthetic-data generation permissions, and creator removal or retraining remedies remain unestablished.

```json
{
  "method": "training-rights-evidence-v1",
  "scope": "DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.",
  "components": {
    "provenance": {
      "score": 2,
      "reason": "The technical report describes a broad data pipeline and training stages, but does not disclose a full source-level corpus and permission ledger for the inherited pretraining mixture.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    },
    "permissions": {
      "score": 0,
      "reason": "A model-weight license governs the released checkpoint, not automatically the training corpus. The reviewed release does not establish affirmative rights for all inherited and post-training material.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    },
    "creatorControls": {
      "score": 0,
      "reason": "The reviewed model documentation does not establish a corpus-wide creator permission, opt-out, removal, or retraining remedy for this model family.",
      "sourceUrls": [
        "https://huggingface.co/deepseek-ai/DeepSeek-V3",
        "https://github.com/deepseek-ai/DeepSeek-V3",
        "https://arxiv.org/abs/2412.19437"
      ]
    }
  },
  "routeDependency": null,
  "modelReference": null
}
```

- [Official model card or model release repository](https://huggingface.co/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Official source repository](https://github.com/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Developer technical report](https://arxiv.org/abs/2412.19437) (primary; checked 2026-10-24)

## Your data & privacy

Finding: Not enough evidence

Assessment: https://ethicalstack.ai/models/deepseek-v3/#privacy

Privacy depends on where these released weights are run; the model family has no single prompt-retention or training route.

Reason: Depends on the selected inference host; do not transfer the model license to a hosting provider’s privacy practices.

Argument: The official release makes weights available for separate deployments, but it does not define one inference host or retention policy for every user. Local and hosted operators may handle prompts differently, so a single numerical privacy grade would misstate this model-level entry.

Type: argument; confidence: high

Scope (profile): DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: The user-selected host, region, telemetry, logging, extensions, and provider training terms are outside the checkpoint release and require a route-specific review.

- [Official model card or model release repository](https://huggingface.co/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Official source repository](https://github.com/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Developer technical report](https://arxiv.org/abs/2412.19437) (primary; checked 2026-10-24)

## Openness

Finding: 8/10

Assessment: https://ethicalstack.ai/models/deepseek-v3/#openness

MIT weights, source code, and detailed architecture/training reports make V3 substantially inspectable and reproducible. Its data sources and permissions remain unresolved; code/weight licensing is not corpus permission.

Reason: MIT for the official V3 repository and weights, as stated in the model card; check V3.2 separately. Model weights and documentation are available, but limitations described above reduce source-level reproducibility.

Argument: MIT weights, source code, and detailed architecture/training reports make V3 substantially inspectable and reproducible. Its data sources and permissions remain unresolved; code/weight licensing is not corpus permission. This openness rating concerns artifacts and reproducibility only; it does not assess whether the original training corpus was authorized.

Type: argument; confidence: medium

Scope (profile): DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: The checked release and license apply to the named checkpoints; sibling variants, future versions, and the underlying corpus may have different terms or disclosures.

- [Official model card or model release repository](https://huggingface.co/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Official source repository](https://github.com/deepseek-ai/DeepSeek-V3) (primary; checked 2026-10-24)

- [Developer technical report](https://arxiv.org/abs/2412.19437) (primary; checked 2026-10-24)

## Worker protections

Finding: 1/10

Assessment: https://ethicalstack.ai/models/deepseek-v3/#labour

DeepSeek’s reviewed model, privacy and user-term materials do not establish employment conditions, a worker grievance route or labor assurance for its AI supply chain.

Reason: 1 base + 0 conditions + 0 worker voice/remedy + 0 evidence/follow-through = 1/10. This grades documented safeguards, not proven mistreatment.

Argument: The sources reviewed provide product, terms, or corporate information but no qualifying disclosure about conditions for data annotators, evaluators, moderators, or infrastructure workers. This is a disclosure gap within the checked materials, not evidence that workers are mistreated or that no labor policy exists.

Type: argument; confidence: low

Scope (dimension): DeepSeek workforce and disclosed suppliers; model-specific coverage may be incomplete.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Worker counts, locations, pay, labor protections, worker voice mechanisms, supplier coverage, and independent audit or remediation outcomes were not established in the sources reviewed.

```json
{
  "method": "worker-safeguards-evidence-v1",
  "entity": "models/deepseek-v3",
  "company": "",
  "scope": "DeepSeek workforce and disclosed suppliers; model-specific coverage may be incomplete.",
  "components": {
    "conditions": {
      "score": 0,
      "reason": "Reviewed sources provide no qualifying disclosure of worker conditions; this zero is limited to the checked material and is not an exhaustive claim.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    },
    "voice": {
      "score": 0,
      "reason": "No qualifying worker-voice mechanism or documented worker grievance outcomes were disclosed in the reviewed material.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    },
    "assurance": {
      "score": 0,
      "reason": "No independent labor audit results, worker-level corroboration, or remediation outcomes were disclosed in the reviewed material.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    }
  },
  "disclosureLimited": true
}
```

- [DeepSeek V4 Public Summary of Training Content](https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf) (primary; checked 2026-10-24)

- [DeepSeek Terms of Use](https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html) (primary; checked 2026-10-24)

- [DeepSeek Privacy Policy](https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html) (primary; checked 2026-10-24)

## Military ties

Finding: 7/10

Assessment: https://ethicalstack.ai/models/deepseek-v3/#military-contracts

No DeepSeek military customer, contract or deployment is identified in the reviewed official sources.

Reason: 7 base + 0 safeguard adjustment = 7/10 (0 of 9 safeguard points; every 3 adds 1). No military relationship was established in the cited sources. This ranks above known defense ties without claiming a verified clean record.

Argument: DeepSeek’s reviewed official terms, privacy material and model disclosures do not identify a defense customer, military contract or deployment. Public model artifacts can be reused downstream, but that does not establish DeepSeek’s own military relationship or any customer-specific safeguard.

Type: argument; confidence: low

Scope (dimension): DeepSeek's relationships and safeguards; this does not identify every product or model used.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Complete contract texts, undisclosed customers, end-use restrictions, actual workload deployment, and independent compliance or outcome reports are unavailable in the reviewed public material.

```json
{
  "method": "military-ties-priority-v2",
  "entity": "models/deepseek-v3",
  "company": "",
  "scope": "DeepSeek's relationships and safeguards; this does not identify every product or model used.",
  "components": {
    "restrictions": {
      "score": 0,
      "reason": "The reviewed terms or policy material contains no qualifying military-use restriction beyond general lawfulness/export compliance; this zero is limited to the reviewed source set.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    },
    "oversight": {
      "score": 0,
      "reason": "No military-specific review, enforcement, or remedy mechanism was disclosed in the sources reviewed.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    },
    "transparency": {
      "score": 0,
      "reason": "No military-specific relationship or military-specific transparency policy was disclosed in the materials reviewed; this is a scoped disclosure finding.",
      "sourceUrls": [
        "https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf",
        "https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html",
        "https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html"
      ]
    }
  },
  "disclosureLimited": true,
  "relationship": "not-established",
  "relationshipReason": "DeepSeek’s model disclosures and user terms reviewed here do not identify a military contract, named defense customer, or military deployment. Its open model artifacts permit downstream uses, but that does not establish DeepSeek’s own military relationship.",
  "score": 7,
  "baseScore": 7,
  "safeguardPoints": 0,
  "safeguardAdjustment": 0,
  "rawScore": 7,
  "cap": null
}
```

- [DeepSeek V4 Public Summary of Training Content](https://cdn.deepseek.com/policies/DeepSeek-V4-public-summary-of-training-content.pdf) (primary; checked 2026-10-24)

- [DeepSeek Terms of Use](https://cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html) (primary; checked 2026-10-24)

- [DeepSeek Privacy Policy](https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html) (primary; checked 2026-10-24)

## BDS targets

Finding: Not listed

Assessment: https://ethicalstack.ai/models/deepseek-v3/#israel-palestine

DeepSeek is not named in the two official BNC campaign lists checked. This is a limited list check, not BDS approval.

Reason: Campaign category, excluded from numerical scoring.

Argument: The official BNC consumer boycott/pressure guide and No Tech campaign page do not name DeepSeek among their current targets. BDS uses selective campaigns, so absence from these lists is neutral and does not establish a clean record or approval.

Type: argument; confidence: medium

Scope (profile): DeepSeek V3 and V3.2 open checkpoints, the released base lineage underlying R1; distinct from the newer DeepSeek V4 profile.

Research reviewed: not recorded; evidence date: not recorded

Unknowns: This check covers only named targets in the two cited BNC campaigns; it does not cover all BDS campaigns, local groups, investments, suppliers, or downstream hosts. Lists and ownership may change.

- [BNC official campaign target list](https://bdsmovement.net/Guide-to-BDS-Boycott) (primary; checked 2026-10-24)

- [BNC official campaign target list](https://bdsmovement.net/no-tech-oppression-apartheid-or-genocide) (primary; checked 2026-10-24)

## Reading the evidence

- Assessments are provisional editorial judgments. Cite the profile and the underlying sources, and retain confidence, scope and unknowns.

- Unknown, not-applicable and model-dependent are different states. Missing evidence does not establish good or bad conduct.

- Company-level evidence does not establish use of a particular product or model. Account plan, model, host and route can change an assessment.

- Energy scores describe documented practices and transparency, not measured request emissions. Zero component points can mean an evidence gap.

- Published overall scores require at least four of six numerical dimensions. Preserve coverage and BDS adjustments; BDS itself is categorical.

- No named BDS target found is limited to the reviewed campaign sources and is not approval. Personalized priorities do not change the published overall score.

- Review and source-check dates describe the research, not this deployment. Fetch current primary sources when the decision requires current verification.

Methodology: https://ethicalstack.ai/methodology/