Registry / No-training default

Which AI providers don't train on your API data?

As of 3 October 2026, of 16 AI model offerings tracked, 14 have clear public evidence and 2 have no public evidence.

Clear public evidence: OpenAI API, Azure OpenAI Service, Anthropic API, Claude via AWS Bedrock, Claude via Google Vertex AI, Gemini via Vertex AI, AWS Bedrock (platform), Mistral La Plateforme, Mistral via Azure AI, Cohere via AWS Bedrock, Llama via AWS Bedrock, Llama via Azure AI, xAI API, DeepSeek via Fireworks AI
No public evidence: Cohere API, DeepSeek API (first-party)

Public commitments not to train on customer API data by default, cell by cell, with sources. The cell answers: Is there a public commitment not to train on customer API data by default? Statuses below are evidence grades, not endorsements, “no public evidence” means we could not verify it from public sources, not that the answer is no.

OpenAI API first-party API

●Yes, public confidence: high · verified 2026-10-03

The OpenAI help article publicly states that by default API data is not used to train models, fulfilling the commitment.

source · full cell

Azure OpenAI Service OpenAI model, served by Microsoft Azure

●Yes, public confidence: high · verified 2026-10-03

Microsoft’s Azure OpenAI data‑privacy page publicly states that prompts and completions are not used to train foundation models without customer permission, confirming a public commitment not to train on customer API data by default.

source · full cell

Anthropic API first-party API

●Yes, public confidence: high · verified 2026-09-29

Anthropic’s own privacy page publicly states that API inputs/outputs are not used for model training by default.

source · full cell

Claude via AWS Bedrock Anthropic model, served by AWS Bedrock

●Yes, public confidence: high · verified 2026-09-30

The AWS Bedrock FAQ publicly commits that customer content is not used to improve base models, confirming a default no‑training‑on‑API‑data policy for Claude via Bedrock.

source · full cell

Claude via Google Vertex AI Anthropic model, served by Google Cloud Vertex AI

●Yes, public confidence: high · verified 2026-10-01

The provider's public data governance page explicitly states a commitment not to use customer data for training by default, and Claude is offered as a managed model on the Gemini Enterprise Agent Platform, so the commitment covers it.

source · full cell

Gemini via Vertex AI Google model, served by Google Cloud Vertex AI

●Yes, public confidence: high · verified 2026-09-30

The provider's own public documentation explicitly states a training restriction, confirming a public commitment not to train on customer API data by default.

source · full cell

AWS Bedrock (platform) platform row

●Yes, public confidence: high · verified 2026-10-02

The AWS Bedrock privacy page publicly states that customer prompts and outputs are not used to train the models by default, confirming the commitment.

source · full cell

Mistral La Plateforme first-party API

●Yes, public confidence: high · verified 2026-09-28

The provider’s own Additional Terms publicly state it does not train on customer data, confirming a default commitment not to train on API data.

source · full cell

Mistral via Azure AI Mistral AI model, served by Microsoft Azure

●Yes, public confidence: high · verified 2026-09-28

The official Azure Foundry FAQ publicly states that customer data is not used to retrain models, confirming the commitment for the Mistral offering.

source · full cell

Cohere API Cohere model, served by Cohere (first-party)

○No public evidence confidence: high · verified 2026-10-03

The provider’s own page states data is used for training by default with an opt‑out toggle, confirming there is no public commitment not to train on customer API data.

source · full cell

Cohere via AWS Bedrock Cohere model, served by AWS Bedrock

●Yes, public confidence: high · verified 2026-10-03

The AWS Bedrock FAQ publicly states that customer content is not used to improve base models, which applies to Cohere models accessed via Bedrock.

source · full cell

Llama via AWS Bedrock Meta model, served by AWS Bedrock

●Yes, public confidence: high · verified 2026-10-01

The AWS Bedrock FAQ publicly states that customer content is not used to improve base models, confirming a default commitment not to train on API data.

source · full cell

Llama via Azure AI Meta model, served by Microsoft Azure (Azure AI Foundry / Models-as-a-Service)

●Yes, public confidence: high · verified 2026-10-01

The Microsoft Foundry FAQ publicly states that customer data is not used to train models, covering Llama offered via Azure AI.

source · full cell

xAI API xAI model, served by xAI (first-party)

●Yes, public confidence: high · verified 2026-10-02

Verbatim quote found on xAI's own docs.x.ai security FAQ, an ungated public commitment not to train on API inputs/outputs without explicit permission, so the recorded yes_public stands.

source · full cell

DeepSeek API (first-party) first-party API

○No public evidence confidence: high · verified 2026-09-29

The provider explicitly notes that some user input may be used for training, contradicting a commitment not to train on API data.

source · full cell

DeepSeek via Fireworks AI DeepSeek model, served by Fireworks AI

●Yes, public confidence: high · verified 2026-10-01

Fireworks AI's public privacy policy explicitly states they do not use API inputs to train models without explicit opt‑in, confirming a default commitment not to train on customer data.

source · full cell