AI Agents & Assistants
Mistral Vibe
EU-Made Report an error0–5 in half steps. 5 means the rubric's top anchor is met on the evidence.
by Mistral AI · mistral.ai
Report an error on this page Is this your product? →
Link to this page
Add ?theme=dark to the image address for dark backgrounds.
The badge shows the category and the month of the latest evaluation, and EU-Made where
that is the product’s provenance. It never shows a score.
Mistral Vibe at a glance
- Vendor
- Mistral AI
- Headquarters
- Not stated on the vendor’s pages as of 26 Oct 2026
- Data hosting
- Not stated on the vendor’s pages as of 26 Oct 2026
- Free plan
- Yes, per the vendor’s own page source, captured 26 Oct 2026
- Panel verdict
- The AI panel rates Mistral Vibe 2.5/5, the mean of 6 judges. Mistral Vibe, the renamed Le Chat from Mistral AI, scores best on pricing transparency and admin and access, both clustering at 5-6: tiers run from a free plan through Pro at $14.99 /mo Excluding taxes to Team at $24.99 /user/mo Excluding taxes, and Team and Enterprise plans document single sign-on, user management with permissions, audit logs, an admin API and data export. how scores are made
- As of
- Vendor pages captured 26 Oct 2026; evaluated by the panel 28 Oct 2026. all sources
Read this page as one judge. Each weighs the same scores by what they care about.
The panel's verdict
Mistral Vibe, the renamed Le Chat from Mistral AI, scores best on pricing transparency and admin and access, both clustering at 5-6: tiers run from a free plan through Pro at $14.99 /mo Excluding taxes to Team at $24.99 /user/mo Excluding taxes, and Team and Enterprise plans document single sign-on, user management with permissions, audit logs, an admin API and data export. Processing and models scores lowest, clustering at 3-4: the catalogue is public, but the sole processing signal is a regional-inference tier label, with no public information on where inference runs. Sovereignty scores sit at a flat 4 — a French company in Paris under French-law SCCs with the CNIL as supervisory authority — yet the only captured hosting fact covers the marketing website on Netlify in San Francisco, and the price list includes Z.ai GLM 5.3 from a non-European maker. Data use and training scores spread from 4 to 6: one side credits the published addendum and per-surface retention; the other weighs default-on training, feedback trained as controller, and zero-data retention unavailable for Vibe Work and Chat.
Speaks for it
- Published pricing from a free tier through Pro at $14.99 /mo Excluding taxes and Team at $24.99 /user/mo Excluding taxes, plus per-model API rates per million tokens with a USD/EUR toggle
- A public data processing addendum names controller and processor roles, states retention per product, and carries French-law standard contractual clauses
- Team and Enterprise plans document single sign-on, user management with permissions, company-wide controls, an admin API, data export and audit logs
- A wide integration surface — 100+ connectors, custom MCP connectors in beta, a coding CLI with hooks and an Agents API — is advertised
- Task scheduling with per-tier limits, workflows, skills, and remote coding agents with observability are documented
Held against it
- Customer input and output train Mistral's models unless opted out, and in-app thumbs feedback is used for training as controller
- Zero-data retention applies only to stateless API calls on pay-as-you-go, requires stated legitimate reasons, is approved at Mistral's discretion, and is unavailable for Vibe Work or Chat regardless of plan
- We found no public information on where inference or customer data is hosted; the only captured hosting fact covers the marketing website on Netlify in San Francisco
- Enterprise deployments are presented as a contact-sales form, with no public information on minimum seats, contract term or overage rates beyond pay-as-you-go credits
- We found no public information on human approval before consequential agent actions, on evaluation or testing before release, or on vendor-stated limits for unsupervised agent work
Best for
- You want a European contracting entity: Mistral AI is a French company in Paris, with a French-law data processing addendum, standard contractual clauses and the CNIL as supervisory authority
- You need to budget seats and API usage before committing, since tier prices, storage limits and per-token rates are public
- Your team wants raw surface to wire assistants into company systems — connectors, a coding CLI with hooks, workflows, an Agents API and an admin API
- Your usage is predominantly stateless API calls on pay-as-you-go, where a zero-data-retention path exists subject to approval at Mistral's discretion
Avoid if
- You must pin the assistant's processing or data hosting to a stated region — the captured pages show only regional-inference tier labels on the API pricing page, with no named region
- Your compliance position requires training to be excluded by contract rather than by opt-out, or zero-data retention on the chat product itself
- You need approval gates or vendor-stated limits before agents act unsupervised on consequential actions in production systems
- You must forecast the enterprise bill precisely before contacting sales, since enterprise deployments present a contact-sales form rather than listed prices
The scores
| Criterion | Panel median |
|---|---|
| Data use, training & retention | 2.5 of 5 |
| Models & processing location | 2 of 5 |
| Admin, identity & access controls | 3 of 5 |
| Integrations, connectors & knowledge | 3 of 5 |
| Agent capabilities & guardrails | 2.5 of 5 |
| European sovereignty | 2 of 5 |
| Pricing transparency | 3 of 5 |
Data use, training & retention
Show reasoningHide reasoning
How this is scored
What happens to prompts, uploaded files, connected data and outputs: whether any of it trains a model (the vendor's or a model provider's), how long it is kept and where logs live, whether zero-retention is available, and whether a data processing agreement covers it.
0 — No statement on training or retention; the privacy policy is a website policy that does not cover the product.
3 — A general statement that business data is "not used for training", with no word on retention, on the model providers behind the product or on logs.
5 — Business data excluded from training by contract, retention periods stated for chats and files, and a DPA offered to business customers.
8 — No-training commitments stated for the vendor and every model provider it routes to, retention configurable by the admin (including deletion of chat history), zero-data-retention terms with the model providers named, and abuse-monitoring retention described.
10 — Data use fully accounted for: training, retention, logging and abuse monitoring stated per model provider, customer-controlled retention and deletion with an audit trail, outputs and connected data treated as customer data in the DPA, and an independent audit (SOC 2, ISO 27001 or ISO 42001) covering the product.
The Head of IT
Retention is documented in real detail — chats kept until the user deletes the conversation or the account, API inputs and outputs held for thirty rolling days for abuse monitoring, fine-tuning data until deleted — and a data processing agreement with controller and processor roles is published. But the default is that inputs and outputs train Mistral's models unless opted out, in-app thumbs feedback is used for training even under business terms, and the vendor states zero-data-retention is unavailable for Vibe Work and Chat regardless of plan, applying only to stateless API calls on pay-as-you-go with approval at Mistral's discretion. I could not sign this without a fight with our data protection officer. 5 6 7
The Data Protection Officer
A data processing agreement is published with Mistral as processor, retention is stated per product — conversations until deletion after a short grace period, API input and output for thirty rolling days for abuse monitoring — and zero-data retention exists for stateless API calls at the vendor's discretion. However training is the default with an opt-out rather than a contractual exclusion, thumbs feedback together with the associated input and output trains Mistral's models as controller, and zero retention is refused outright for the chat and agent products regardless of plan. We found no public information on no-training commitments covering the third-party models in the lineup. 2 5 6 7 12
The Automation Lead
The data processing addendum puts Mistral in the processor role and states retention precisely per surface — until conversation or account deletion for the assistant, thirty rolling days for API abuse monitoring — with a zero-data-retention path, but the vendor states it applies only to stateless API calls and not to the chat product, agents or conversations, at the vendor's discretion. Model training runs on customer input and output unless the customer opts out, and admin-configurable retention is not described in anything I read. 2 5 6 7
The Customer-Service Director
The data processing agreement names us Controller and Mistral Processor, and retention is unusually explicit — Vibe conversations kept until we delete them, API traffic thirty rolling days for abuse monitoring. But inputs and outputs train Mistral's models unless we opt out, and zero-data-retention is stated as unavailable for Vibe, Chat, agents and conversations, applying only to stateless API calls at Mistral's discretion — so every chat transcript from my contact centre would sit in the default retention. 2 5 6 7
The Works Council Member
Training on staff input is the default, with an opt-out per account, and thumbs up/down feedback hands the associated input and output to model training even for business customers. Retention is stated plainly — chats kept until the user deletes them, API input and output thirty rolling days for abuse monitoring — but zero-data-retention is documented as unavailable for the chat product itself regardless of subscription plan. A public, detailed data processing agreement under French law with standard contractual clauses is genuinely more than most vendors publish. 2 5 6 7
The Skeptic
Retention is documented per product — Vibe conversations kept until deleted, API inputs and outputs held thirty rolling days for abuse monitoring, fine-tuning data until you delete it — and a data processing addendum with SCCs and breach notification is offered. But training is opt-out rather than excluded: Mistral trains on customer input and output unless opted out, and thumbs up/down feedback always trains as controller. Zero-data-retention exists only for stateless API calls on pay-as-you-go, must be justified, is approved "at Mistral's discretion", and is explicitly unavailable for Vibe Work or Chat. 5 6 7
Models & processing location
Show reasoningHide reasoning
How this is scored
Which models sit behind the product and who runs them, where inference and storage happen, whether the customer can pin processing to the EU, choose or exclude model providers, bring its own model or key, or run the product in its own cloud or on premises.
0 — The models behind the product and the processing location are not stated.
3 — The model family is named, but where inference runs and through which provider (the model maker, a hyperscaler, the vendor) is unstated.
5 — Models and their providers named, with an EU processing option stated for storage and at least the default model.
8 — Each available model listed with provider and inference region, EU-only processing selectable by the admin and enforced across models, model providers excludable, and bring-your-own-key or own-cloud deployment documented.
10 — Processing fully under the customer's control and documented: EU-only or on-premises inference for every feature including agents and connectors, open-weight or self-hosted models supported, data flows per feature published, and a residency commitment in the contract rather than on a marketing page.
The Head of IT
The model families are named in depth — Mistral Large, Medium, Small, Ministral, Codestral, Voxtral, OCR and a self-hosted open-weights Shieldstral whose costs depend on our own infrastructure — plus a third-party Z.ai GLM on the API price list. Where inference runs and through which provider is not stated, the only hint being regional-inference Standard, Batch and Priority tiers on the API page, and we found no public information on an admin-selectable EU processing option. For an 800-person deployment I cannot pin processing from these pages. 10 12
The Data Protection Officer
The model lineup is public and priced, from Ministral and Mistral Large through Codestral, Voxtral and OCR to a third-party Z.ai family, alongside a self-hosted open-weights option whose costs depend on the customer's own infrastructure. The pricing pages mention regional inference tiers without a stated region, and we found no public information on an EU processing option for the assistant itself, on excluding model providers, or on bring-your-own-key. 2 10 12
The Automation Lead
The model families are named across the pricing and models pages, with a regional-inference tier selector on the API pricing page, a self-hosted open-weights product and custom deployments on team plans — but we found no public information on where inference and storage for the assistant itself run, or any EU processing commitment for the product. 2 9 10 12
The Customer-Service Director
The full model catalogue is public with per-token prices, including third-party Z.ai GLM models and one self-hosted open-weights option, but we found no public information on where inference runs or any admin setting to pin processing to the EU — the enterprise deployments page is a contact-sales form rather than a specification. 4 9 10 12
The Works Council Member
The model catalogue is published down to per-token prices, but we found no public information on where inference runs or through which provider. A "regional inference" toggle appears on the API pricing page without a named region, so an admin cannot pin the assistant to EU processing on the strength of anything we read. 10 12
The Skeptic
The model families are named and priced — Mistral Large, Medium, Small, Ministral, Voxtral, plus Z.ai GLM 5.3 from a non-European maker — and a self-hosted open-weights model and Partner-Served deployment terms exist. But the only word on processing location is a "Regional inference" tier label on the API pricing page; we found no public information on inference regions, an EU processing commitment for Vibe, or bring-your-own-key. 6 10 12
Admin, identity & access controls
Show reasoningHide reasoning
How this is scored
What the IT department controls: SSO and SCIM provisioning, roles and groups, which models, agents and connectors each group may use, retrieval that respects the permissions of the source system, audit logs, usage reporting, and policies such as data-loss prevention — and whether reporting on staff use is scoped and documented.
0 — No admin features documented; accounts are individual sign-ups.
3 — A team workspace with an owner and member invitations, but no SSO, no roles and no audit log.
5 — SSO, admin and member roles, user management, and an admin view of usage.
8 — SAML/OIDC SSO with SCIM provisioning, group-based permissions for models, agents and connectors, retrieval that inherits source-system permissions, exportable audit logs, and usage analytics with documented scope.
10 — Governance a regulated company can operate: fine-grained policies (DLP, content filters, prompt and output logging rules), audit logs streamed to a SIEM, per-agent ownership and approval, usage reporting that can be aggregated or pseudonymised for works-council agreements, and every admin action logged.
The Head of IT
Single sign-on and SAML are documented for Team and Enterprise, alongside user management with permissions, company-wide admin controls, an admin API, and audit logs listed among team and enterprise features. We found no public information on SCIM provisioning, group-based entitlements to models, agents or connectors, retrieval that respects source-system permissions, or export of audit logs to a SIEM. That covers the sign-in basics but not the governance layer this rollout needs. 11 2
The Data Protection Officer
Team and Enterprise plans document SAML single sign-on, user management with permissions, company-wide administrator controls, an admin API, data export and audit logs. We found no public information on SCIM provisioning, on group-level restrictions governing which models, agents or connectors each group may use, or on retrieval that respects the permissions of the source system. 11 2
The Automation Lead
Team and Enterprise plans get SAML single sign-on, user management with permissions, company-wide controls such as an organization-wide Flash Answers opt-out, audit logs and an Admin API — a real admin surface. We found no public information on SCIM provisioning, group-level control over which models or connectors each team may use, or retrieval that inherits the permissions of the source system. 2 11
The Customer-Service Director
Team and Enterprise plans document an admin panel with single sign-on, user management with permissions, company-wide controls, audit logs and an admin API. We found no public information on SCIM provisioning, group-level permissions over connectors, or retrieval that respects the permissions of the source system — the last matters when the assistant reads customer records. 2 7 11
The Works Council Member
Single sign-on, user management with permissions and audit logs are documented for Team and Enterprise plans, plus an admin API and data export. We found no public information on usage reporting that could be aggregated or pseudonymised, or on prompt-logging rules set out in writing — the two things a works council needs before staff use this. 2 11
The Skeptic
Team and Enterprise administrators get SAML SSO, user management with permissions, company-wide controls, an admin API, audit logs and data export. We found no public information on SCIM provisioning, group-based permissions over models, agents or connectors, or the scope and exportability of audit logs and usage reporting. 2 11
Integrations, connectors & knowledge
Show reasoningHide reasoning
How this is scored
How the product reaches company knowledge and systems: connectors to file stores, wikis, mail, CRM, ticketing and ERP, how fast and how completely they sync, the API, support for open protocols such as MCP, and how the assistant cites the sources it answers from.
0 — No integrations; users paste text or upload files one at a time.
3 — File upload and one or two connectors (for example a drive), with no API and no statement on sync or citations.
5 — A connector library covering the common file stores, wikis and collaboration tools, answers citing sources, and a documented API.
8 — Connectors to file stores, wikis, mail, CRM and ticketing with incremental sync, custom connectors or MCP support, an API and webhooks for embedding assistants and agents in other tools, and source citations on every grounded answer.
10 — The product sits in the company's systems: read and write connectors with documented scopes, MCP servers and clients, contact-centre or telephony integration where relevant, a knowledge layer that can be audited and corrected, and the data, agents and configuration exportable when the customer leaves.
The Head of IT
The product advertises over one hundred connectors plus custom MCP connectors in beta, libraries and document upload with per-user storage limits, project folders, and a coding CLI with connectors, hooks and remote agents. We found no public information on which systems the connectors cover, how fast or completely they sync, or how answers cite their sources — permission-aware SharePoint retrieval is the make-or-break for us, and these pages do not show it. A token-priced API is documented. 2 12
The Data Protection Officer
The product lists one hundred-plus connectors, custom MCP connectors in beta, document libraries with per-tier storage limits, workflows and an admin API, and the API side is priced per model with per-token, per-thousand-page and per-minute rates. We found no public information on sync speed or completeness, on webhooks, or on how answers cite the sources they are grounded in. 2 4 12
The Automation Lead
The pricing pages claim 100+ connectors with custom MCP connectors and knowledge connectors for teams, plus an Admin API, a CLI and hooks — the raw surface I need to wire agents into company systems. The connectors are labelled beta, the captured pages do not show which systems they reach or how they sync, and we found no public information on source citations or webhooks. 2 7 12
The Customer-Service Director
A hundred-plus connectors, custom MCP connectors in beta and knowledge connectors on Team and Enterprise, plus a documented API, give real reach into company knowledge. We found no public information on connectors into ticketing or CRM specifically, on sync frequency, or on how the assistant cites the sources behind an answer — exactly the systems a contact centre lives in. 2 12
The Works Council Member
Over a hundred connectors, custom MCP connectors in beta, and document libraries up to 30GB per user are listed on the pricing page, with model and admin APIs priced and named. We found no public information on how connectors sync, or on whether answers cite the sources they were built from. 2 12
The Skeptic
A connector library of "100+ connectors" is advertised (listed as beta), custom MCP connectors are in beta, libraries and document upload come with stated storage limits, and an admin API and Connector Terms exist. We found no public information on sync behaviour, connector scopes, or how grounded answers cite their sources. 2 4
Agent capabilities & guardrails
Show reasoningHide reasoning
How this is scored
What an agent can actually do and where it stops: multi-step tasks, tool use and actions in other systems, scheduling and triggers, human approval before consequential actions, run logs, evaluation and testing before release, and the limits the vendor states rather than hides.
0 — No agents; the product is a chat window, whatever the marketing says.
3 — Saved prompts or custom assistants with instructions and files, but no actions in other systems and no run history.
5 — Agents that call tools or APIs and run multi-step tasks, built without code, with a log of each run.
8 — Agents that act in connected systems with scoped permissions, human approval configurable per action, triggers and schedules, test and evaluation tooling before release, and documented limits and failure handling.
10 — Agents a business can put in production and account for: versioning and staged rollout, evaluation against test sets, monitoring of quality and cost per agent, handover to a human with full context, every action attributable to the agent and the person who deployed it, and the vendor's own statement of what agents must not be trusted with.
The Head of IT
Task scheduling with per-tier limits, skills, workflows, and a remote coding agent mode with observability are in the feature list, and an Agents API exists with its retention stated. We found no public information on human approval gates before consequential actions, evaluation or testing before release, or the limits the vendor states for what agents must not be trusted with. Impressive assistants, but not yet agents I can put in production and account for. 2 5
The Data Protection Officer
Task scheduling with tier-based limits, workflows, skills, MCP tools and remote coding agents with agent observability are documented, and an Agents API exists with stated retention of input and output until account termination. We found no public information on human approval before consequential actions, on evaluation or testing before release, on run logs beyond the remote coding agents, or on vendor-stated limits on what agents should not be trusted with. 2 5 12
The Automation Lead
There are real agents here — task scheduling, workflows, remote coding agents with observability, a voice agents API billed per minute, and an Agents API whose inputs and outputs are kept until account termination — so multi-step work with a trace of runs is documented. We found no public information on human approval before consequential actions, scoped per-agent permissions, or evaluation and testing before release; without approval gates I would not put these in front of the ERP. 2 5 12
The Customer-Service Director
Task scheduling, workflows, skills, remote agent observability and a Voice Agents API billed per minute of conversation show genuine multi-step automation without code. But we found no public information on agents acting inside order or ticket systems, on human approval before consequential actions, on handover to a human with full context, or on evaluation and testing before release — all four are make-or-break before I put an agent in front of customers. 2 5 12
The Works Council Member
Workflows, task scheduling with unlimited runs on paid tiers, and remote coding agents with observability are named as features. We found no public information on human approval before an agent takes a consequential action, on run logs a council could inspect, or on the vendor stating what agents must not be trusted with. 2
The Skeptic
Task scheduling with per-tier limits, workflows, an Agents API, hooks and remote coding agents with observability go beyond saved prompts — observability at least implies run visibility, and scheduling is a trigger. We found no public information on human approval before consequential actions, evaluation or testing before release, or the limits the vendor states for what agents may do unsupervised. 2 5
European sovereignty
panel opinion
Show reasoningHide reasoning
How this is scored
Who the contracting entity and the parent company are, where customer data is hosted, and who the subprocessors are — model providers above all, then hosting, search, telephony and email. Independently sourced by the sovereignty pipeline; for an AI product the model provider is a subprocessor, so a European interface over a non-European model counts as such.
0 — Non-EU vendor and contracting entity, hosting unstated, subprocessors (including model providers) unnamed.
3 — A privacy policy exists, but hosting location is unstated or non-EU and the model providers behind the product are not named as subprocessors.
5 — EU hosting stated and an EU contracting entity, but prompts reach non-EU model providers or other subprocessors without a stated safeguard or EU region.
8 — EU hosting on named infrastructure, EU contracting entity, every subprocessor listed with its location including model providers, EU inference available, and a DPA offered.
10 — Sovereign end to end and evidenced: European ownership, entity, hosting and model processing, every subprocessor European or optional, data export and deletion documented, and a security or certification statement published rather than implied.
The Head of IT
The contracting entity is French — Mistral AI, registered in Paris, with a Paris data protection officer and the CNIL named as supervisory authority, French governing law in the standard contractual clauses, and subprocessors audited under signed data protection agreements. But we found no public information on where the product hosts customer data — the only hosting detail the captured pages give is for the marketing website, on Netlify in San Francisco — and the privacy policy acknowledges transfers outside the EU under Article 46 safeguards. A European vendor I can contract with, but residency for my data is not proven here. 3 5 6
The Data Protection Officer
Mistral AI is a French company contracting in France with the CNIL as supervisory authority, a French-law data processing agreement with standard contractual clauses, and a stated practice of auditing every customer-data provider and signing a dedicated data protection agreement with each. But the only hosting fact captured concerns the marketing website, hosted by Netlify in San Francisco — we found no public information on where the product's customer data is hosted, no subprocessor list with locations, and no region statement for the Z.ai models appearing in the lineup. 3 5 6 12
The Automation Lead
The contracting entity is proven European — Mistral AI, registered in Paris under number 952 418 325, with French law in the data processing addendum, the CNIL as supervisory authority — and providers are audited and bound by their own data protection agreements before handling customer data. The only hosting fact captured covers the marketing website (Netlify in San Francisco), and the captured pages do not state where customer data for the product is hosted or name the subprocessors. 3 5 6 7
The Customer-Service Director
Mistral is a French company in Paris under French law, with the CNIL as supervisory authority and Standard Contractual Clauses attached to contracts with providers outside the EU — a strong European base. We found no public information on where customer data is hosted or on the named subprocessors with their locations, only that each provider is audited and a data protection agreement signed. 3 5 6
The Works Council Member
The contracting entity is a French company in Paris under French law, with CNIL named as the supervisory authority and standard contractual clauses in the data processing agreement. We found no public information on where customer data for the assistant is hosted or on a named subprocessor list with locations — a Trust Center is referenced, but its contents are not in the pages we read. 5 6 7
The Skeptic
The contracting entity is proven European — Mistral AI, a French company in Paris, French-law SCCs, CNIL as supervisory authority — which is real. But we found no public information on where Vibe customer data is hosted, subprocessors are audited and contracted but unnamed, transfers outside the EU are possible, and the API catalog routes to Z.ai GLM 5.3, a non-European model. 3 5 6 12
Pricing transparency
Show reasoningHide reasoning
How this is scored
Whether a company can compute the real annual cost from public pages alone: price per seat, per agent, per conversation or per usage unit, what is included in each tier, usage limits and overage, which models, connectors or agents cost extra, minimum seats and term.
0 — No public prices at all; every customer is a sales call.
3 — A starting price or "from" figure exists, but usage limits, included models and what scales the cost are unclear.
5 — Prices per tier are public with the main inclusions, but usage limits, overage, agent runs or enterprise features (SSO, audit logs, EU processing) are unpriced or "on request".
8 — Every tier priced publicly with seat or usage units, usage limits and overage rates, add-ons listed, minimum seats and term stated.
10 — Complete price computability: annual cost derivable for a given number of seats, agents or conversations, with every usage rate, model surcharge, add-on and support tier and the cancellation and data-return terms published.
The Head of IT
Free, Pro at "$14.99 /mo" and Team at "$24.99 /user/mo" are published excluding taxes, with tier limits such as unlimited task scheduling on paid plans, per-user storage, and an API page that prices every model per million tokens in USD or EUR. Enterprise deployments are a contact-sales form, minimum seats and term are unstated, and overage beyond pay-as-you-go credits is unpriced for the assistant itself. I can budget the API; I cannot compute the enterprise bill. 2 9 12
The Data Protection Officer
The tiers are public — a free plan with stated limits, "$14.99 /mo Excluding taxes" for Pro and "$24.99 /user/mo Excluding taxes" for Team with storage and scheduling limits per tier — and the API is priced per million tokens with cached rates, per-thousand-page OCR, per-minute voice and a USD/EUR toggle. Enterprise deployments are presented as a contact-sales form rather than listed prices, message limits are expressed as multiples of the free plan, and we found no public information on overage rates beyond pay-as-you-go credits, minimum seats or contract terms. 2 9 12
The Automation Lead
Free, Pro at "$14.99 /mo Excluding taxes" and Team at "$24.99 /user/mo Excluding taxes" are published with storage and scheduling limits per tier, and the API page prices every model per million tokens, per thousand pages or per minute with a USD/EUR toggle. The enterprise tier is a contact-sales form, and we found no public information on pay-as-you-go overage rates for the assistant, minimum seats or contract terms. 2 9 12
The Customer-Service Director
Tiers are public with stated limits — $14.99 /mo Pro and $24.99 /user/mo Team, both excluding taxes, with caps on messages, scheduled tasks and storage — and the API page prices every model per million tokens, including voice conversations at $0.04/min for Mistral Small 4 and $0.08/min for larger models. The enterprise tier is a contact-sales form with no listed prices, and we found no public information on overage rates for the assistant or on minimum seats and term, so the CFO can forecast seats and voice minutes but not the full annual contract. 2 9 12
The Works Council Member
Published prices run from a free tier through $14.99/mo and $24.99/user/mo excluding taxes, with message, search and storage limits stated per tier, and the API priced per million tokens down to cached inputs. Enterprise pricing is a contact-sales form, and we found no public information on overage rates, minimum seats or contract term. 2 9 12
The Skeptic
Pro at $14.99 /mo and Team at $24.99 /user/mo, excluding taxes, are public with main inclusions, and the API page unit-prices every model per million tokens, per minute and per thousand pages with a USD/EUR toggle — that side is computable. Enterprise deployments are a contact-sales form, tier usage limits are expressed as multiples like "Up to 6x free" rather than absolute numbers, and we found no public information on minimum seats, term or overage rates beyond pay-as-you-go credits. 2 9 12
European sovereignty — proven facts
0 of 4 dimensions provenBuilt only from facts shown on the vendor's own pages. A dimension we could not prove is left open, not scored as zero.
| Legal entity | Not determined | — | uncited Report an error |
|---|---|---|---|
| Ownership | Not determined | — | uncited Report an error |
| Data residency | Not determined | — | uncited Report an error |
| Subprocessors | Not determined | — | uncited Report an error |
Questions about Mistral Vibe
Is Mistral Vibe European?
Yes. whats-best.ai classifies Mistral Vibe as EU-Made.
Where does Mistral Vibe store data?
Not stated on the vendor’s pages as of 26 Oct 2026.
Does Mistral Vibe have a free plan?
Yes. Mistral Vibe has a free plan, per the vendor’s page captured 26 Oct 2026.
What are the best alternatives to Mistral Vibe?
As of 28 Oct 2026, the panel rates Langdock (4/5), Dust (3.5/5) and Unique AI (3.5/5) highest among the other 15 AI Agents & Assistants products evaluated. Ratings that look equal are rounded to half stars; the order uses the unrounded rating, then the name. The highest-rated European alternative is Langdock (EU-Made). Ratings are the AI panel’s opinion, formed from public evidence: how scores are made. The full list, with European products first: Mistral Vibe alternatives.
Where this could be wrong
- Evidence ages. The oldest capture behind this page is from 26 Oct 2026. Vendors change pricing and policies without notice; every fact reflects its source as of the capture date shown in the registry.
- AI can misread a source. Extraction and judgement are automated; a citation guarantees traceability, not infallibility. If something here is wrong, say so — no account needed, every report is decided within 5 business days, and accepted corrections are published.
What we left out
A claim that does not survive our checks costs us the claim, not the page. This is what was taken off this one.
- 24 product facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 10 pricing facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 8 compliance facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 3 integrations facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 2 hosting facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 2 legal facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 1 data fact could not be confirmed on the vendor’s page as captured and was left out of this page and of the panel’s material. Know more? Tell us
- 1 support fact could not be confirmed on the vendor’s page as captured and was left out of this page and of the panel’s material. Know more? Tell us
Sources (12)
The pages every claim on this page was read from — each one checked, dated, and kept verifiable.
- 1 Vendor homepage (Vibe, formerly Le Chat) mistral.ai Checked 26 Oct 2026 Details →
- 2 Pricing page mistral.ai Checked 26 Oct 2026 Details →
- 3 Legal notice mistral.ai Checked 26 Oct 2026 Details →
- 4 Legal center with terms of service legal.mistral.ai Checked 26 Oct 2026 Details →
- 5 Privacy policy legal.mistral.ai Checked 26 Oct 2026 Details →
- 6 Data processing addendum legal.mistral.ai Checked 26 Oct 2026 Details →
- 7 Data use, training & retention — found from sitemap help.mistral.ai Checked 26 Oct 2026 Details →
- 8 Data use, training & retention — found from sitemap help.mistral.ai Checked 26 Oct 2026 Details →
- 9 Models & processing location — found from sitemap mistral.ai Checked 26 Oct 2026 Details →
- 10 Models & processing location — found from sitemap help.mistral.ai Checked 26 Oct 2026 Details →
- 11 Admin, identity & access controls — found from sitemap help.mistral.ai Checked 26 Oct 2026 Details →
- 12 Integrations, connectors & knowledge — found from sitemap mistral.ai Checked 26 Oct 2026 Details →