Claude vs. DeepSeek: data protection per plan
Claude (6 plans) and DeepSeek (2 plans) side by side: DPA, training on inputs, hosting and data retention per plan, based on the providers’ public documents. Without rating.
Last updated: (latest review date of the tools shown)
At a glance
Each value is listed with the plans it applies to. Pick a tier to see only plans of that tier; if plans are chosen in the plan comparison below, the table shows exactly those.
All plans: each value is listed with the plans it applies to.
| Feature | Claude | DeepSeek |
|---|---|---|
| Provider based in | Anthropic PBC: 548 Market St, PMB 90375, San Francisco, CA 94104, USA | Hangzhou DeepSeek Artificial Intelligence Co., Ltd., 'registered address in China' |
| Plans recorded | 6 | 2 |
| Hosting | mixed: Free, Pro, Max, API / Consolenot documented: Team, Enterprise | not documented: all 2 plans |
| Training on inputs | opt-out available: Free, Pro, Maxno: Team, Enterprise, API / Console | opt-out available: Chatnot documented: Open Platform API |
| DPA | not documented: Free, Pro, Maxyes: Team, Enterprise, API / Console | not documented: all 2 plans |
| As of | 18 Oct 2026 | 18 Oct 2026 |
Plans compared
Comparing: Claude Free vs. DeepSeek Chat
| Feature | ClaudeFree | DeepSeekChat (web and app) |
|---|---|---|
| not publicly documented | not publicly documented | |
| Details: DPA available | not publicly documented (the Consumer Terms and Privacy Policy do not mention a DPA; according to the Privacy Policy, Anthropic Ireland Limited is the controller for users in the EEA/UK/CH) | not publicly documented |
| opt-out available | opt-out available | |
| Details: Training on inputs | Consumer Terms (Effective 2025-10-08): "We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review". Privacy Center (2026-03-16): chats and coding sessions are used if users allow it, in the case of safety flagging or with an express opt-in. Incognito chats: "are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings". The pricing page shows "Model training: Opt-out" for Free, Pro and Max. According to the Privacy Center, this also applies to Claude Code for accounts on these plans. | Terms of use (as of 2026-03-27, section 4.3): DeepSeek may use inputs and outputs 'to a minimal extent', after encryption and de-identification, to improve the services and the underlying technologies; opt-out available by switching off 'Improve the model for everyone'. Privacy policy: data is used, among other things, to train and improve the models; right to object to use for training. Document 'Model Mechanism and Training Methods of DeepSeek': in the optimisation phase, a small portion of the training data may be based on user input, which is then encrypted, de-identified and anonymised. |
| differs by plan | mixed | not publicly documented |
| Details: Data residency | Privacy Policy (Effective 2026-09-10): "your personal data is transferred to our servers in the US, or to other countries outside the European Economic Area ('EEA') and the UK"; Trust Center: cloud infrastructure (Google Cloud, AWS, Microsoft Azure) with location "Worldwide" for all products. A choice of region is not documented for consumer plans. | Privacy policy (as of 2026-02-10): 'we directly collect, process and store your Personal Data in People's Republic of China'. The category 'US only/EU option' does not apply; the selection list does not include China, hence 'not publicly documented' as the enum value here. An EU or US region: not publicly documented. |
| differs by plan | list available | not publicly documented |
| Details: List published |
| not publicly documented |
| differs by plan | Deleted chats: removed from the history… | Privacy policy: storage for as long as… |
| Details: Data retention | Deleted chats: removed from the history immediately, deleted from the back-end systems "within 30 days". If training is allowed: de-identified retention for up to 5 years in training pipelines (only for new/resumed chats after activation). In the case of usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01). | Privacy policy: storage for as long as necessary for the services and the stated purposes; for accounts, as long as the account exists; legal and legitimate interests may justify longer retention. Fixed retention periods: not publicly documented. Chat history can be deleted in the settings; after account deletion, access cannot be restored. |
| differs by plan | 0 USD/mo | free |
| Details: Price | USD 0/month (list claude.com/pricing, as of 2026-10-14; prices excl. taxes) | free of charge (website app description: 'Chat with leading AI models for free.'; no paid chat plan found) |
| yes | yes | |
| Details: AI features included | yes (chat, web search, file creation, code execution, Memory, Artifacts, Connectors; according to the pricing page) | yes (AI chatbot; according to the terms of use, services based on large language models) |
| 6 sources | 4 sources | |
| Details: Sources with retrieval date |
|
|
Contract / DPA
- Claude Freenot publicly documented
- DeepSeek Chatnot publicly documented
- Claude Free
not publicly documented (the Consumer Terms and Privacy Policy do not mention a DPA; according to the Privacy Policy, Anthropic Ireland Limited is the controller for users in the EEA/UK/CH)
DeepSeek Chatnot publicly documented
Training
- Claude Freeopt-out available
- DeepSeek Chatopt-out available
- Claude Free
Consumer Terms (Effective 2025-10-08): "We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review". Privacy Center (2026-03-16): chats and coding sessions are used if users allow it, in the case of safety flagging or with an express opt-in. Incognito chats: "are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings". The pricing page shows "Model training: Opt-out" for Free, Pro and Max. According to the Privacy Center, this also applies to Claude Code for accounts on these plans.
DeepSeek ChatTerms of use (as of 2026-03-27, section 4.3): DeepSeek may use inputs and outputs 'to a minimal extent', after encryption and de-identification, to improve the services and the underlying technologies; opt-out available by switching off 'Improve the model for everyone'. Privacy policy: data is used, among other things, to train and improve the models; right to object to use for training. Document 'Model Mechanism and Training Methods of DeepSeek': in the optimisation phase, a small portion of the training data may be based on user input, which is then encrypted, de-identified and anonymised.
Hosting
- differs by plan
- Claude Freemixed
- DeepSeek Chatnot publicly documented
- Claude Free
Privacy Policy (Effective 2026-09-10): "your personal data is transferred to our servers in the US, or to other countries outside the European Economic Area ('EEA') and the UK"; Trust Center: cloud infrastructure (Google Cloud, AWS, Microsoft Azure) with location "Worldwide" for all products. A choice of region is not documented for consumer plans.
DeepSeek ChatPrivacy policy (as of 2026-02-10): 'we directly collect, process and store your Personal Data in People's Republic of China'. The category 'US only/EU option' does not apply; the selection list does not include China, hence 'not publicly documented' as the enum value here. An EU or US region: not publicly documented.
Sub-processors
- differs by plan
- Claude Freelist available
- DeepSeek Chatnot publicly documented
- Claude Free
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- Persona (fraud detection, identity verification; USA; Free/Pro/Max only)
- Yoti (fraud detection, identity verification; UK; Free/Pro/Max only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- Note: according to the Privacy Policy, Anthropic is the controller for consumers; the list nevertheless assigns these service providers to the products Claude Free/Pro/Max.
DeepSeek Chatnot publicly documented
Retention
- differs by plan
- Claude FreeDeleted chats: removed from the history…
- DeepSeek ChatPrivacy policy: storage for as long as…
- Claude FreeDeleted chats: removed from the history immediately, deleted from the back-end systems "within 30 days". If training is allowed: de-identified retention for up to 5 years in training pipelines (only for new/resumed chats after activation). In the case of usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01).DeepSeek ChatPrivacy policy: storage for as long as necessary for the services and the stated purposes; for accounts, as long as the account exists; legal and legitimate interests may justify longer retention. Fixed retention periods: not publicly documented. Chat history can be deleted in the settings; after account deletion, access cannot be restored.
Price
- differs by plan
- Claude Free0 USD/mo
- DeepSeek Chatfree
- Claude Free
USD 0/month (list claude.com/pricing, as of 2026-10-14; prices excl. taxes)
DeepSeek Chatfree of charge (website app description: 'Chat with leading AI models for free.'; no paid chat plan found)
- Claude Freeyes
- DeepSeek Chatyes
- Claude Free
yes (chat, web search, file creation, code execution, Memory, Artifacts, Connectors; according to the pricing page)
DeepSeek Chatyes (AI chatbot; according to the terms of use, services based on large language models)
Evidence
- Claude Free6 sources
- DeepSeek Chat4 sources
- Claude Free
- Anthropic Privacy Center: Is my data used for model training? (Consumer: Free, Pro, Max; as of 2026-03-16)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: How long do you store my data? (Consumer; as of 2026-07-01)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Consumer Terms of Service (Effective 2025-10-08)Terms and conditions/Terms, retrieved 18 Oct 2026
- Anthropic Privacy Policy (Effective 2026-09-10)Privacy notice, retrieved 18 Oct 2026
- Claude Pricing (Individual, Team & Enterprise, API)Pricing page, retrieved 18 Oct 2026
- Anthropic Trust Center: Subprocessors (22 entries)Sub-processor list, retrieved 18 Oct 2026
DeepSeek Chat- DeepSeek Privacy Policy (Last Update 2026-02-10)Privacy notice, retrieved 18 Oct 2026
- DeepSeek Terms of Use (Last Update 2026-03-27)Terms and conditions, retrieved 18 Oct 2026
- Model Mechanism and Training Methods of DeepSeekDocumentation, retrieved 18 Oct 2026
- DeepSeek website (home page and Transparency Center)Product page, retrieved 18 Oct 2026
| Feature | Claude | DeepSeek |
|---|---|---|
| differs by tool |
| not publicly documentedall 2 plans |
| Details: DPA available |
|
|
| differs by tool |
|
|
| Details: Training on inputs |
|
|
| differs by tool |
| not publicly documentedall 2 plans |
| Details: Data residency |
|
|
| differs by tool | list availableall 6 plans | not publicly documentedall 2 plans |
| Details: List published |
|
|
| differs by tool |
| Privacy policy: storage for as long as…all 2 plans |
| Details: Data retention |
|
|
| differs by tool |
|
|
| Details: Price |
|
|
| yesall 6 plans | yesall 2 plans | |
| Details: AI features included |
|
|
| 17 sources | 6 sources | |
| Details: Sources with retrieval date |
|
|
Contract / DPA
- differs by tool
- Claude
- not publicly documentedFree, Pro, Max
- yesTeam, Enterprise, API / Console (Claude Platform)
- DeepSeeknot publicly documentedall 2 plans
- Claude
- Free
not publicly documented (the Consumer Terms and Privacy Policy do not mention a DPA; according to the Privacy Policy, Anthropic Ireland Limited is the controller for users in the EEA/UK/CH)
- Pro
not publicly documented (the Consumer Terms and Privacy Policy do not mention a DPA; according to the Privacy Policy, Anthropic Ireland Limited is the controller for users in the EEA/UK/CH)
- Max
not publicly documented (the Consumer Terms and Privacy Policy do not mention a DPA; according to the Privacy Policy, Anthropic Ireland Limited is the controller for users in the EEA/UK/CH)
- Team
yes: Commercial Terms section C: "Data submitted through the Services will be processed in accordance with the Anthropic Data Processing Addendum ('DPA'), which is incorporated into these Terms by reference." (Anthropic Ireland Limited is the contracting party where the customer is established in the EEA, Switzerland or the UK)
Provider’s DPA - Enterprise
yes: Commercial Terms section C: "Data submitted through the Services will be processed in accordance with the Anthropic Data Processing Addendum ('DPA'), which is incorporated into these Terms by reference." (Anthropic Ireland Limited is the contracting party where the customer is established in the EEA, Switzerland or the UK)
Provider’s DPA - API / Console (Claude Platform)
yes: Commercial Terms section C: "Data submitted through the Services will be processed in accordance with the Anthropic Data Processing Addendum ('DPA'), which is incorporated into these Terms by reference." (Anthropic Ireland Limited is the contracting party where the customer is established in the EEA, Switzerland or the UK)
Provider’s DPA
DeepSeek- Chat (web and app)
not publicly documented
- Open Platform API (usage-based)
not publicly documented
- Free
Training
- differs by tool
- Claude
- opt-out availableFree, Pro, Max
- noTeam, Enterprise, API / Console (Claude Platform)
- DeepSeek
- opt-out availableChat (web and app)
- not publicly documentedOpen Platform API (usage-based)
- Claude
- Free
Consumer Terms (Effective 2025-10-08): "We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review". Privacy Center (2026-03-16): chats and coding sessions are used if users allow it, in the case of safety flagging or with an express opt-in. Incognito chats: "are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings". The pricing page shows "Model training: Opt-out" for Free, Pro and Max. According to the Privacy Center, this also applies to Claude Code for accounts on these plans.
- Pro
Consumer Terms (Effective 2025-10-08): "We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review". Privacy Center (2026-03-16): chats and coding sessions are used if users allow it, in the case of safety flagging or with an express opt-in. Incognito chats: "are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings". The pricing page shows "Model training: Opt-out" for Free, Pro and Max. According to the Privacy Center, this also applies to Claude Code for accounts on these plans.
- Max
Consumer Terms (Effective 2025-10-08): "We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review". Privacy Center (2026-03-16): chats and coding sessions are used if users allow it, in the case of safety flagging or with an express opt-in. Incognito chats: "are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings". The pricing page shows "Model training: Opt-out" for Free, Pro and Max. According to the Privacy Center, this also applies to Claude Code for accounts on these plans.
- Team
Privacy Center (2026-08-18): "By default, we will not use your inputs or outputs from our commercial products (e.g. Claude for Work, Anthropic API, Claude Gov, etc.) to train our models." Exceptions according to the same source: express feedback (thumbs up/down) or other express permission. Commercial Terms: "Anthropic may not train models on Customer Content from Services." Pricing page Team: "No model training on your content by default".
- Enterprise
Privacy Center (2026-08-18): "By default, we will not use your inputs or outputs from our commercial products (e.g. Claude for Work, Anthropic API, Claude Gov, etc.) to train our models." Exceptions according to the same source: express feedback (thumbs up/down) or other express permission. Commercial Terms: "Anthropic may not train models on Customer Content from Services." Pricing page Enterprise model training: "None by default".
- API / Console (Claude Platform)
Privacy Center (2026-08-18): "By default, we will not use your inputs or outputs from our commercial products (e.g. Claude for Work, Anthropic API, Claude Gov, etc.) to train our models." Exceptions according to the same source: express feedback (thumbs up/down) or other express permission. Commercial Terms: "Anthropic may not train models on Customer Content from Services."
DeepSeek- Chat (web and app)
Terms of use (as of 2026-03-27, section 4.3): DeepSeek may use inputs and outputs 'to a minimal extent', after encryption and de-identification, to improve the services and the underlying technologies; opt-out available by switching off 'Improve the model for everyone'. Privacy policy: data is used, among other things, to train and improve the models; right to object to use for training. Document 'Model Mechanism and Training Methods of DeepSeek': in the optimisation phase, a small portion of the training data may be based on user input, which is then encrypted, de-identified and anonymised.
- Open Platform API (usage-based)
The Open Platform terms (effective 2026-04-29) govern ownership of inputs and outputs (the customer retains rights to inputs and receives rights to outputs) but contain no express statement on the use of API inputs for training; they refer to the document 'Model Principles and Training Methods Explanation', which likewise does not distinguish between chat and API. The terms of use (section 4.3) and the privacy policy apply to 'Services' in general; an API-specific opt-out is not described. For end-user data in its own applications, according to the privacy policy and the Open Platform terms (5.5), the developer is the controller.
- Free
Hosting
- differs by tool
- Claude
- mixedFree, Pro, Max, API / Console (Claude Platform)
- not publicly documentedTeam, Enterprise
- DeepSeeknot publicly documentedall 2 plans
- Claude
- Free
Privacy Policy (Effective 2026-09-10): "your personal data is transferred to our servers in the US, or to other countries outside the European Economic Area ('EEA') and the UK"; Trust Center: cloud infrastructure (Google Cloud, AWS, Microsoft Azure) with location "Worldwide" for all products. A choice of region is not documented for consumer plans.
- Pro
Privacy Policy (Effective 2026-09-10): "your personal data is transferred to our servers in the US, or to other countries outside the European Economic Area ('EEA') and the UK"; Trust Center: cloud infrastructure (Google Cloud, AWS, Microsoft Azure) with location "Worldwide" for all products. A choice of region is not documented for consumer plans.
- Max
Privacy Policy (Effective 2026-09-10): "your personal data is transferred to our servers in the US, or to other countries outside the European Economic Area ('EEA') and the UK"; Trust Center: cloud infrastructure (Google Cloud, AWS, Microsoft Azure) with location "Worldwide" for all products. A choice of region is not documented for consumer plans.
- Team
not publicly documented for claude.ai workspaces (Team/Enterprise): no choice of region documented. The Trust Center states the location "Worldwide" for the cloud infrastructure (Google Cloud, AWS, Microsoft Azure); Privacy Policy: transfer "to our servers in the US, or to other countries outside the [EEA] and the UK".
- Enterprise
not publicly documented for claude.ai workspaces (Team/Enterprise): no choice of region documented. The Trust Center states the location "Worldwide" for the cloud infrastructure (Google Cloud, AWS, Microsoft Azure); Privacy Policy: transfer "to our servers in the US, or to other countries outside the [EEA] and the UK". The Claude Enterprise product page (q19) lists "US-only inference" among the data controls; an EU region is not mentioned there.
- API / Console (Claude Platform)
Claude Platform Docs (Data residency): two settings. Inference geo (parameter inference_geo): "global" (default) "Inference may run in any available geography"; "us" "Inference runs only in US-based infrastructure" (surcharge 1.1x, Claude 4.6 and newer only). Workspace geo (storage at rest and endpoint processing): "Currently, 'us' is the only available workspace geo." According to the same documentation (retrieved 2026-10-18): "Inference geo: Only 'us' and 'global' are available." and "Workspace geo: Only 'us' is currently available."; an EU region is therefore not currently offered. On Amazon Bedrock/Google Cloud, the endpoint URL or the inference profile determines the region (not part of this entry).
DeepSeek- Chat (web and app)
Privacy policy (as of 2026-02-10): 'we directly collect, process and store your Personal Data in People's Republic of China'. The category 'US only/EU option' does not apply; the selection list does not include China, hence 'not publicly documented' as the enum value here. An EU or US region: not publicly documented.
- Open Platform API (usage-based)
Privacy policy (as of 2026-02-10): 'we directly collect, process and store your Personal Data in People's Republic of China'. The category 'US only/EU option' does not apply; the selection list does not include China, hence 'not publicly documented' as the enum value here. An EU or US region: not publicly documented. Region of API processing: not publicly documented.
- Free
Sub-processors
- differs by tool
- Claudelist availableall 6 plans
- DeepSeeknot publicly documentedall 2 plans
- Claude
- Free
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- Persona (fraud detection, identity verification; USA; Free/Pro/Max only)
- Yoti (fraud detection, identity verification; UK; Free/Pro/Max only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- Note: according to the Privacy Policy, Anthropic is the controller for consumers; the list nevertheless assigns these service providers to the products Claude Free/Pro/Max.
- Pro
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Stripe (billing; USA)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- Persona (fraud detection, identity verification; USA; Free/Pro/Max only)
- Yoti (fraud detection, identity verification; UK; Free/Pro/Max only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- Note: according to the Privacy Policy, Anthropic is the controller for consumers; the list nevertheless assigns these service providers to the products Claude Free/Pro/Max.
- Max
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Stripe (billing; USA)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- Persona (fraud detection, identity verification; USA; Free/Pro/Max only)
- Yoti (fraud detection, identity verification; UK; Free/Pro/Max only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- Note: according to the Privacy Policy, Anthropic is the controller for consumers; the list nevertheless assigns these service providers to the products Claude Free/Pro/Max.
- Team
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Stripe (billing; USA)
- WorkOS (security, single sign-on; USA)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- ElevenLabs (text-to-speech; USA; Claude for Work only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- Enterprise
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Stripe (billing; USA)
- WorkOS (security, single sign-on; USA)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- ElevenLabs (text-to-speech; USA; Claude for Work only)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
- API / Console (Claude Platform)
- According to the Trust Center (trust.anthropic.com/subprocessors, 22 entries, retrieved 2026-10-18):
- Google Cloud Platform (cloud infrastructure; worldwide; all products)
- Amazon Web Services (cloud infrastructure; worldwide; all products)
- Microsoft Azure (cloud infrastructure; worldwide; all products)
- Cloudflare (traffic routing/CDN; worldwide, close to the customer; all products)
- Stripe (billing; USA)
- WorkOS (security, single sign-on; USA)
- Intercom (user support; USA)
- Nutun (user support; South Africa)
- Boldr (user support; Canada)
- Twilio (analytics, e-mail/SMS; USA)
- Iterable (e-mail communication; USA)
- Functional Software dba Sentry (error handling, support; USA)
- Sift (fraud/abuse detection; USA)
- Arkose Labs (fraud/abuse detection; USA)
- Brave Search (web search; USA)
- TurboPuffer (web search; USA)
- Intuition Machines, Inc. (hCaptcha) (bot detection, fraud/abuse detection; worldwide; all products except Claude for Government)
- Parallel Web Systems, Inc. (web search; USA; all products except Claude for Government)
DeepSeek- Chat (web and app)not publicly documented
- Open Platform API (usage-based)not publicly documented
- Free
Retention
- differs by tool
- Claude
- Deleted chats: removed from the history…Free, Pro, Max
- Saved chats/coding sessions are retained…Team, Enterprise
- Anthropic API: inputs/outputs are…API / Console (Claude Platform)
- DeepSeekPrivacy policy: storage for as long as…all 2 plans
- Claude
- FreeDeleted chats: removed from the history immediately, deleted from the back-end systems "within 30 days". If training is allowed: de-identified retention for up to 5 years in training pipelines (only for new/resumed chats after activation). In the case of usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01).
- ProDeleted chats: removed from the history immediately, deleted from the back-end systems "within 30 days". If training is allowed: de-identified retention for up to 5 years in training pipelines (only for new/resumed chats after activation). In the case of usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01).
- MaxDeleted chats: removed from the history immediately, deleted from the back-end systems "within 30 days". If training is allowed: de-identified retention for up to 5 years in training pipelines (only for new/resumed chats after activation). In the case of usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01).
- TeamSaved chats/coding sessions are retained in the product for continuing the conversation; deleted chats: removed from the history immediately, deleted from the back end "within 30 days"; incognito chats are automatically deleted within 30 days. Usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01). Help Center (q18, as of 2026-03-16): "While using Team and Enterprise plans, prompts and conversations in the product are maintained so you can see your conversation history in the product."
- EnterpriseSaved chats/coding sessions are retained in the product for continuing the conversation; deleted chats: removed from the history immediately, deleted from the back end "within 30 days"; incognito chats are automatically deleted within 30 days. Usage policy violation flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback data: 5 years (Privacy Center, as of 2026-07-01). Custom data retention controls (Enterprise): the default is unlimited retention; organisation owners can set the retention period for chats and projects, minimum 30 days (Help Center); according to the Help Center, the custom periods do not apply to Claude Design, Claude Tag, Claude Managed Agents and other features built on Claude Code on the web. According to the Help Center, zero data retention agreements apply to eligible APIs and Claude Code on Enterprise plans, not to the entire product. Help Center (q18): "Enterprise plan customers can also set custom retention timelines for their organization's data."
- API / Console (Claude Platform)Anthropic API: inputs/outputs are automatically deleted in the back end "within 30 days of receipt or generation", except for services with longer retention under customer control (e.g. Files API), an agreed deviation (e.g. Zero Data Retention), usage policy enforcement or a legal obligation. Limited retention and review apply to Covered Models. Flagging: inputs/outputs up to 2 years, classification scores up to 7 years. Feedback: 5 years. ZDR possible on application and after approval (Help Center, as of 2026-07-01).
DeepSeek- Chat (web and app)Privacy policy: storage for as long as necessary for the services and the stated purposes; for accounts, as long as the account exists; legal and legitimate interests may justify longer retention. Fixed retention periods: not publicly documented. Chat history can be deleted in the settings; after account deletion, access cannot be restored.
- Open Platform API (usage-based)Privacy policy: storage for as long as necessary for the services and the stated purposes; for accounts, as long as the account exists; legal and legitimate interests may justify longer retention. Fixed retention periods: not publicly documented. API-specific retention: not publicly documented.
Price
- differs by tool
- Claude
- 0 USD/moFree
- 17 USD/mo… (further prices in the details)Pro
- 100 USD/mo… (further prices in the details)Max
- several tiersTeam
- 20 USD/user/moEnterprise
- per tokenAPI / Console (Claude Platform)
- DeepSeek
- freeChat (web and app)
- per tokenOpen Platform API (usage-based)
- Claude
- Free
USD 0/month (list claude.com/pricing, as of 2026-10-14; prices excl. taxes)
- Pro
USD 17/month with annual subscription (USD 200 in advance), USD 20/month with monthly billing (list claude.com/pricing, as of 2026-10-14; excl. taxes)
- Max
Max 5x: USD 100/month; Max 20x: USD 200/month; prices for web subscriptions; according to the pricing page FAQ, annual subscriptions only for Pro and Team ("Annual subscriptions are available for Pro and Team plans.") (Help Center "What is the Max plan?", q17; pricing page claude.com/pricing: "from USD 100", prices excl. taxes)
- Team
Standard seat USD 20/seat/month with annual billing, USD 25 with monthly billing; Premium seat USD 100/seat/month annually, USD 125 monthly; Team for 2 to 150 people (list claude.com/pricing, as of 2026-10-14; excl. taxes)
- Enterprise
Self-serve: USD 20/seat/month with annual billing, minimum 20 seats, plus usage-based costs at API rates; sales-assisted: individual quote (pricing page FAQ: "You can get started self-serve or contact sales for a tailored quote."); Enterprise with annual billing only (list claude.com/pricing, retrieved 2026-10-18)
- API / Console (Claude Platform)
List claude.com/pricing (as of 2026-10-14), per 1M tokens: Fable 5.1 input USD 10 / output USD 50; Opus 5.5 input USD 4 / output USD 20; Sonnet 5.5 input USD 2 / output USD 10; Haiku 5.5 input USD 0.10 (prompts up to 100K) / output USD 0.50; Batch -50 %; US-only inference at 1.1x price (excl. taxes)
DeepSeek- Chat (web and app)
free of charge (website app description: 'Chat with leading AI models for free.'; no paid chat plan found)
- Open Platform API (usage-based)
Per 1M tokens, in USD (pricing page): deepseek-flash: input (cache hit) USD 0.003 off-peak / USD 0.006 peak; input (cache miss) USD 0.15 / USD 0.3; output USD 0.6 / USD 1.2. deepseek-v4-pro: cache hit USD 0.022 / USD 0.044; cache miss USD 0.66 / USD 1.32; output USD 1.98 / USD 3.96. Peak hours: 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday (excluding Chinese public holidays); otherwise half price. Search and Fetch API USD 0.75 each per 1,000 successful requests. Billed against topped-up or granted credit; DeepSeek may change prices; no indication whether incl. VAT.
- Free
- Claudeyesall 6 plans
- DeepSeekyesall 2 plans
- Claude
- Free
yes (chat, web search, file creation, code execution, Memory, Artifacts, Connectors; according to the pricing page)
- Pro
yes (Free plus more usage, Claude Code, Claude Design/Slides/Docs, Research, Claude in Chrome and Microsoft 365; according to the pricing page)
- Max
yes (Pro plus 5x or 20x more usage, higher output limits, Early Access, Priority Access; new: monthly API credits for the Claude Platform of USD 100 (Max 5x) and USD 200 (Max 20x) respectively; according to the pricing page)
- Team
yes (all Claude features, more usage than Pro, Claude Code and Claude Cowork, Enterprise Search, SSO; according to the pricing page up to USD 500/month in pooled API credits)
- Enterprise
yes (all Team features plus roles/SCIM, audit logs, Compliance API, IP allowlisting, HIPAA-ready offering; according to the pricing page)
- API / Console (Claude Platform)
yes (Messages API, Batch, Managed Agents and others; prices per MTok)
DeepSeek- Chat (web and app)
yes (AI chatbot; according to the terms of use, services based on large language models)
- Open Platform API (usage-based)
yes (models deepseek-flash = DeepSeek-V4.1-Flash-1017 and deepseek-v4-pro = DeepSeek-V4-Pro-0813, each with 1M-token context, up to 384K output, thinking/non-thinking mode, tool calls, JSON output, Responses API, OpenAI- and Anthropic-compatible interface; vision only with deepseek-flash; additionally Search and Fetch API)
- Free
Evidence
- Claude17 sources
- DeepSeek6 sources
- Claude
- Anthropic Privacy Center: Is my data used for model training? (Consumer: Free, Pro, Max; as of 2026-03-16)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: Is my data used for model training? (Commercial; as of 2026-08-18)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: How long do you store my data? (Consumer; as of 2026-07-01)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: How long do you store my organization's data? (Commercial; as of 2026-07-01)Privacy notice/Help Center, retrieved 18 Oct 2026
- Anthropic Data Processing Addendum (Effective 2025-02-24)DPA, retrieved 14 Oct 2026
- Anthropic Commercial Terms of Service (Effective 2025-06-17)Terms and conditions/Terms, retrieved 14 Oct 2026
- Anthropic Consumer Terms of Service (Effective 2025-10-08)Terms and conditions/Terms, retrieved 18 Oct 2026
- Anthropic Privacy Policy (Effective 2026-09-10)Privacy notice, retrieved 18 Oct 2026
- Claude Pricing (Individual, Team & Enterprise, API)Pricing page, retrieved 18 Oct 2026
- Claude Platform Docs: Data residencyDocumentation, retrieved 18 Oct 2026
- Anthropic Privacy Center: Configure custom data retention controls for Enterprise plansHelp Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: Zero data retention agreement - applicable productsHelp Center, retrieved 14 Oct 2026
- Anthropic Privacy Center: How do you use personal data in model training? (Commercial; as of 2026-03-16)Privacy notice/Help Center, retrieved 14 Oct 2026
- Anthropic Trust Center: Subprocessors (22 entries)Sub-processor list, retrieved 18 Oct 2026
- Claude Help Center: What is the Max plan?Help Center, retrieved 18 Oct 2026
- Anthropic Privacy Center: Can you delete data that I sent via Team and Enterprise plans? (as of 2026-03-16)Privacy notice/Help Center, retrieved 18 Oct 2026
- Claude Enterprise (product page incl. FAQ Security and compliance)Product page, retrieved 18 Oct 2026
DeepSeek- Models & PricingPricing page, retrieved 18 Oct 2026
- DeepSeek Privacy Policy (Last Update 2026-02-10)Privacy notice, retrieved 18 Oct 2026
- DeepSeek Terms of Use (Last Update 2026-03-27)Terms and conditions, retrieved 18 Oct 2026
- DeepSeek Open Platform Terms of Service (effective 2026-04-29)Terms and conditions, retrieved 18 Oct 2026
- Model Mechanism and Training Methods of DeepSeekDocumentation, retrieved 18 Oct 2026
- DeepSeek website (home page and Transparency Center)Product page, retrieved 18 Oct 2026
not documentedmeans: not publicly documented, not verifiable in any provider source. It does not mean “no”.
No guarantee of accuracy, not legal advice. Each entry is valid as of the date shown. Open points are listed on the detail pages.
More comparisons
Claude compared with
Individual pages: Claude, DeepSeek. All tools in the category: AI assistants & search.