Mistral trains on user input by default for Vibe, except on the enterprise tier

Mistral's help center spells out that the company trains its models on user input and output data by default for its consumer Vibe product, with users required to opt out manually if they don't want their interactions used for training.

What's new

Mistral's own support article on the policy states plainly: "Vibe: users are not opted out by default and can opt out at any time in their settings. Vibe (Enterprise): customers are opted out of training by default." That means the default posture flips entirely depending on which tier a user is on — consumer Vibe users are opted in unless they act, while Enterprise Vibe customers are opted out unless they choose otherwise.

The article also draws a distinction that's easy to miss: opting out of training on Vibe does not automatically opt a user out on Mistral's separate API and Studio products. "Vibe and API opt-out toggles are separate. Opting out of one does not opt you out of the other — you must configure each toggle individually."

Mistral documents separate opt-out paths for each surface:

• Vibe (web): Admin panel → Vibe → disable "Allow your interactions to be used to train our models"
• Vibe (mobile, iOS/Android): Settings → Data & Account Controls → deselect "Enable data sharing"
• Mistral Studio/API: Admin panel → Privacy → disable "Anonymous improvement data"

Context

Mistral has built its public identity partly around being a European, more transparency-forward alternative to the US frontier labs, and it already publishes granular help-center documentation on data retention and training use across its product lines — a level of specificity not every lab matches. The company draws a clear line between its consumer-facing Vibe assistant, where opt-out is the user's responsibility, and its paid Enterprise and API tiers, where non-training is the default and forms part of the commercial pitch to business customers.

Why it matters

Default settings are what most users actually experience, regardless of what a privacy policy says is technically possible to change. A default-in training posture on a free consumer product is a common industry pattern — several major chat assistants train on free-tier conversations unless a user opts out — but it puts the burden on individual users to find and flip a toggle across web and mobile separately, and again on API access if they use both surfaces. For developers and businesses evaluating Mistral against Enterprise-tier competitors that default to no-training, the split matters directly: only Mistral's paid Enterprise and API tiers get non-training by default, while anyone using the free Vibe consumer product is training data unless they act. That distinction is worth knowing before assuming a Mistral account is privacy-safe by default.