mistral.ai Review: Is It Legit and Safe?
⚠ Policy changed 2026-06-18 — see Policy Changes below for what's different.
Quick verdict: mistral.ai earns a trust grade B, 78.4 out of 100, based on 4 of 5 dimensions.
78.4/100
Grade B · 4 of 5 checks · 1 not yet assessed
Our verdict
mistral.ai earns a grade B (78.4/100) on our evidence-based trust rubric, measured on 4 of 5 trust dimensions. It scores strongly on company & identity but only moderately on data & privacy (5.5/10). We have not yet assessed product reality, so the grade is provisional.
Is mistral.ai reputable?
MixedMostly — it's a real company, but with gaps.
Mistral.ai has a real footprint: the team behind it isn't publicly identified, and 89 Trustpilot reviewers rate it 2.4/5. Some company details we look for are missing or unverified — see the section below before trusting it.
Is mistral.ai safe to use?
YesMostly — the privacy basics check out.
mistral.ai's privacy policy commits to GDPR. The catch: deleting your data needs an email request rather than one click. No billing complaints have been filed with us, but our reports database is still new — check recent Trustpilot reviews on cancellation and refunds before you subscribe.
A plain-language read of the documented evidence below — every claim here is backed by the sources in the sections that follow.
Trust Score: 78.4/100 — Grade B
Scored 2026-08-01 on 4 of 5 trust dimensions (the rest are marked “Not yet assessed” — the grade reflects only what we measured). Every score below is backed by at least one piece of documented evidence — see how we score.
| Dimension | Score | Bar | Source |
|---|---|---|---|
| Company & Identity | 8 / 8 | Automated signals | |
| Data & Privacy | 5.5 / 10 | Automated signals | |
| Billing & Subscription | 8 / 8 | Automated signals | |
| Product Reality | Not yet assessed | ||
| Community Reputation | 4.76 / 7.5 | Automated signals | |
mistral.ai pricing at a glance
From mistral.ai's public pricing page · captured 2026-07-08. Prices change — verify current pricing →
✅ Free tier available
- FreeFree
- Pro$14.99 / month
- Team$24.99 / month
- EnterpriseContact us
- Education plan$5.99
Who's Behind mistral.ai?
WHOIS: mistral.ai registered 2019-05-15 (~7 year(s) old). Real contact information verified (registrant org / site contact page).
What Happens to Your Data
Account/data deletion option per policy: partial (email request required). Compliance commitments in policy — GDPR: yes, CCPA: no. Privacy policy changed on 2026-06-18 (a major change) — Privacy policy for mistral.ai changed: 225 clause(s) added, 14 removed. Flagged as a major change. Notable: "Data publicly available on the Internet. Our artificial intelligence models are trained on data that is publicly available on the Internet by third parties, whi...".
Does mistral.ai train AI on your data?
We don't reduce this to a yes/no — wording matters (defaults, opt-outs, enterprise carve-outs). Here is exactly what mistral.ai's privacy policy said when we captured it on 2026-06-18:
“When you subscribe to a paid Mistral AI Product “Input”, any data you provide to generate an Output, such as prompts or fine-tuning data. “Feedback”, any information you provide when you rate an Output, such as “thumbs up” or “thumbs down", and the associated Input and Output.”
“Our artificial intelligence models are trained on data that is publicly available on the Internet by third parties, which may contain personal data, even if we use good practices to filter out such personal data.”
“We do not use this Third-Party Content to train our models.”
Read it in context: mistral.ai privacy policy ↗ (captured 2026-06-18). If this policy changes materially, it will appear on our policy-change alerts.
What else mistral.ai's policy says
How long they keep your data
“We store your personal data only as long as needed for the purposes described in Section 3 of this Privacy Policy.”
Quoted verbatim from the mistral.ai privacy policy ↗ (captured 2026-06-18). We don't interpret these — read them in context.
Pricing & Cancellation Reality
Specific prices shown publicly per their pricing page. 0 approved billing complaint(s) (charged without consent / cannot cancel) in our reports database as of 2026-08-01.
Product reality
We assess tools from public records, so this grade reflects mistral.ai's company, data, billing and reputation — not how well its output actually performs, which we haven't scored yet. To judge that, try the free tier if there is one and read recent hands-on reviews before committing to a paid plan.
Community Reputation
Trustpilot rating 2.4/5 (89 reviews). App Store (iOS) rating: 4.76/5 from 1,196 ratings.
Official app check: which mistral.ai app is real?
✅ The official iOS app is Vibe by Mistral (ex-Le Chat) by MISTRAL AI (4.76/5 from 1,196 ratings). View on the App Store ↗. Before installing, always check that the developer name matches.
Technical signals for mistral.ai
- HTTPS certificate:Valid — issued by Google Trust Services, expires 2026-07-14
- HSTS header:Strict-Transport-Security enforced
- Security headers:Content-Security-Policy + X-Content-Type-Options set
- Mail server:MX records configured on the domain
- Threat blocklists:Not on the Cloudflare malware blocklist
Network-level facts collected automatically on 2026-06-12. They are not part of the trust score — a missing header is common even on legitimate sites. A flagged threat blocklist is the only red flag in this list.
Frequently Asked Questions
Is mistral.ai legit?
mistral.ai is operated by a registered company, and scores strongly on company & identity in our checks.
Is mistral.ai safe to use?
On data & privacy mistral.ai scores moderately (5.5/10). See the data section above for the specifics we verified.
Is mistral.ai a scam?
No — mistral.ai is a legitimate, registered service (grade B). Note any billing caveats above.
mistral.ai Alternatives
Similar AI tools we track (by semantic similarity), with their trust grades:
Stigg 2.0 The usage runtime for AI products
A
DevCleaner Free the gigabytes your dev tools and AI apps hoard
B
TryCase Disposable test environments for AI coding agents
BCompare mistral.ai head-to-head on trust: