Le Chat
← Back to All AI Tools1. Should you try Le Chat? Overview
| What it is | Le Chat → Mistral Vibe — EU-Jurisdiction AI Assistant With Open-Weight Backbone, Vibe Coding & 256K Context (Mistral AI) Not shut down — rebranded. |
| Country / HQ | Paris, France |
| Best for | EU/EEA companies with GDPR/data-residency mandates (health/finance/government) |
| Check before buying | Watch out: US-only startups indifferent to jurisdiction (ChatGPT/Claude cheaper ecosystem) |
| Editor rating | 4.2 / 5 |
| Pricing signal | Free tier + paid from $5.99/mo |
2. Features What it does
| Feature | What it does |
|---|---|
| Models | Mistral Small 4 (119B MoE, Apache-2.0), Large 3 (256K, MRL license), Medium 3.5 (256K, 77.6% SWE-Bench Verified), Magistral (CoT), Devstral (coding), Voxtral TTS (4B open) — Le Chat/Vibe routes Small 4 free, Large 3/Medium 3.5 paid |
| Chat | Flash Answers, Think mode, web search + citations, doc upload, 500 memories (free) / more (paid), Projects folders — Free ~25 msg/day soft cap |
| Canvas | Doc/table/slide/code/presentation editor in-chat — Pro+ |
| Media | Flux Ultra image gen (free+paid, 40× free on Pro), code interpreter sandbox, voice mode (record/upload→transcribe) — No native video gen (vs Sora) |
| Agents | Work Mode: multi-step mail/calendar/inbox workflows, sensitive actions need explicit approval; Vibe for Code: CLI + IDE + remote sandbox, hooks, teleport, run history/audit — Pro = all-day coding; Free = limited sessions |
| Connectors | 100+ (Notion/Slack/Linear/Drive/GitHub/Outlook/Asana/Snowflake/Stripe/Zapier), custom MCP connectors — All tiers get connectors, rate differs |
| Platform | EU-hosted (GDPR, EU AI Act compliance docs, Aug 2 2026 full enforcement), open weights downloadable, Enterprise on-prem/private cloud, SAML SSO/audit/white-label — Pro = no-telemetry opt-out |
3. Pricing Pricing (2026-10, USD)
| Plan | Price / mo | Key limits & who it fits |
|---|---|---|
| Free | $0 | $0 Small 4, ~25 msg/day soft cap, image gen (limited), 100+ connectors, 500 memories, limited Vibe coding sessions, no no-telemetry |
| Pro | $14.99 | — (no annual discount shown on page) 6× free msgs, 5× web search, 40× image, 15GB storage, Think/deep research, all-day Vibe coding (CLI/IDE/remote), no-telemetry mode, chat+email support |
| Student | $5.99 | — Verified .edu, first-time only, 12-mo validity, Pro features |
| Team | $24.99/seat | — $50/mo floor (≥2 seats), 30GB/seat, domain verify, admin API, data export, shared RAG |
| Enterprise | Custom (~$20K+/mo) | — On-prem/private cloud, custom models, SAML SSO, audit logs, white-label, dedicated CSM |
| API (La Plateforme) | PAYG | — Small 4 $0.15/$0.60 · Codestral $0.30/$0.90 · Large 3 $0.50/$1.50 · Medium 3.5 $1.50/$7.50 per M tok · Pro sub does NOT include API credits |
4. Integration Access & integrations
| Access type | Details |
|---|---|
| Access | access_type: web-app-plus-mobile-plus-cli-plus-ide-plus-api access_display: 🌐 chat.mistral.ai (Vibe web) · 📱 iOS/Android (Vibe) · 💻 Vibe CLI (Devstral) + VS Code/JetBrains plugin · 🔌 La Plateforme REST API + custom MCP connectors · no desktop Electro |
5. Reviews & Sources External reviews
| Source | Rating | What it says |
|---|---|---|
| Stackaible | 4.2/5 — “EU-boundary seat done right; free tier (top models+Flux+connectors) embarrasses ChatGPT free; Pro $14.99 cheapest major-lab paid” | |
| ClickUp Alt Roundup | “multilingual FR/DE/ES native-level, Flash Answers snappy; ecosystem/integrations younger than US trio” | |
| ToolChase | “Work Mode approval-gated agent = safer than Operator; Vibe coding credible vs Claude Code for Devstral-class tasks; 25-msg free cap arrives fast” | |
| AIToolTier | “Small 4 Apache-2.0 + Medium 3.5 77.6% SWE-Bench = real exit ramp; peak reasoning trails GPT-5.6/Claude Opus 4.8; mobile app thinner” |
6. Best for / Not for Fit
| ✔ Suitable | ✘ Not suitable |
|---|---|
| • EU/EEA companies with GDPR/data-residency mandates (health/finance/government) • Multilingual teams (FR/DE/ES/IT/PT) drafting cross-border docs • Cost-sensitive Pro users wanting coding bundled • Devs wanting Devstral coding + open-weight self-host exit • Students (verified .edu $5.99) doing coursework + code • Orgs hedging US-provider concentration risk |
• US-only startups indifferent to jurisdiction (ChatGPT/Claude cheaper ecosystem) • Max-frontier reasoning needs (Claude Opus 4.8/GPT-5.6 Sol) • Voice/video creators (no Sora/advanced voice, Flux only images) • GPTs/Operator/agent-store dependent workflows • Heavy 500K+ token single-doc analysis (Claude 1M/Gemini 1M) • Non-European language-heavy (JA/ZH) — US models stronger |
7. Competitors How it compares
| Competitor | How it differs | Best for |
|---|---|---|
| ChatGPT | US jurisdiction, GPT-5.6, GPTs/Operator/Sora, voice | $0 / $20-mo |
| Claude | Best prose/reasoning, 1M ctx, Claude Code | $0 / $20-mo |
| Gemini | Google-stack, 1M ctx, Veo video | $0 / $19.99-mo |
| Perplexity | Cited search-first, no coding agent/Work Mode | $0 / $20-mo |
| DeepSeek | Cheapest API, open weights, no EU-hosted assistant | $0 / API $0.14/0.28 |
| Grok | X-firehose realtime, Imagine video | $0 / $30-mo |
| Qwen/GLM | Open-weight alt, weaker EU posture | $0 / API low |
| OpenRouter chat | Multi-model router, no EU jurisdiction guarantee | pay-per-tok |
8. Summary Verdict
| Who should pick it | EU/EEA companies with GDPR/data-residency mandates (health/finance/government) |
| Who should skip it | US-only startups indifferent to jurisdiction (ChatGPT/Claude cheaper ecosystem) |
Event history
- 2026-05-28 Rebranded — The product was not shut down, merely rebranded. Accounts, conversation history and subscriptions were all migrated automatically, and the URL `chat.mistral.ai` remains unchanged. The rebrand on **28 May 2026** coincided with a product upgrade: Work Mode, Code Mode, VS Code extension and CLI agent capabilities were added, transforming it beyond a basic chatbot. The former name Le Chat has been officially retired; the public‑facing product is now called Vibe.