Le Chat

Le Chat → Mistral Vibe — EU-Jurisdiction AI Assistant With Open-Weight Backbone, Vibe Coding & 256K Context (Mistral AI)

Not shut down — rebranded.​ Mistral AI (Paris, founded Apr 2023 by Mensch/Lample/Lacroix) renamed Le Chat → Vibe on 2026-05-28; all conversations, settings, plans, and billing carried over automatically, chat.mistral.ai still works, old Le Chat slug redirects to Vibe. This is a rename event, not a shutdown​ — treat le-chat as alias of mistral-vibe in your directory (cross-link, don’t double _event_log). Category ai-productivity-data; pipeline_stage none. Killer differentiator = EU/GDPR-first jurisdiction (data processed under EU law, open-weight models auditable/self-hostable) + Pro $14.99/mo undercuts ChatGPT Plus/Claude Pro by $5 while bundling Vibe coding (Devstral CLI/IDE remote agents) they charge extra for + 256K ctx on Large 3/Medium 3.5 + Flux image gen + 100+ connectors + Work Mode agentic approval-gated actions; not ChatGPT (US jurisdiction, no open weights, $20), not Claude (best prose/reasoning but no EU self-host moat), not Perplexity (search-first, no coding agent/Work Mode), not DeepSeek (cheaper API but no EU-hosted assistant surface). 2026 trap: free tier ~25 msg/day soft cap dies fast, Pro sub does NOT include API credits (two separate bills), Medium 3.5 256K still half Gemini/Claude 1M, no native GPTs/Operator/Sora-class voice-video ecosystem, peak reasoning trails GPT-5.6/Claude Opus 4.8, student $5.99 requires verified .edu + first-time only 12-mo, Team $50/mo floor = 2 seats minimum.

1. One-line positioning

Mistral Vibe (ex-Le Chat) = the EU-boundary assistant: chat.mistral.ai → pick Small 4 (free) / Large 3 / Medium 3.5 → Flash Answers (~1,000 tok/s), Think mode (Magistral CoT), web search with citations, document upload (PDF/CSV/Office), Canvas (doc/table/slide/code), Flux image gen, 100+ connectors (Notion/Slack/Linear/Drive/GitHub/Outlook) → Work Mode runs multi-step mail/calendar/inbox triage with explicit approval gates → Vibe for Code (Devstral-powered CLI + VS Code/JetBrains + remote sandbox agents, teleport session local↔remote) → same weights callable via La Plateforme API (Small 4 $0.15/$0.60, Large 3 $0.50/$1.50, Medium 3.5 $1.50/$7.50 per M). Fits ai-productivity-data; pipeline_stage none (chat→search→reason→act→code→deploy). For EU/GDPR-bound teams, multilingual (FR/DE/ES/IT) work, cost-sensitive Pro users, and devs wanting Devstral coding + self-host exit ramp, it’s the jurisdiction play; for max frontier reasoning/voice/video/GPTs ecosystem, ChatGPT/Claude win.

2. Core features

3. Pricing (2026-08, USD — chat sub ≠ API)

pricing_type = freemium-per-seat-plus-payg-api; starting_price = $0; free_tier = "yes; Small 4, ~25 msg/day soft cap, Flux image gen limited, 100+ connectors, 500 memories, doc upload, code interpreter, limited Vibe coding sessions, no no-telemetry/no deep research/all-day coding"; pricing_note = "le-chat-renamed-vibe-2026-05-28-not-shutdown-carryover-automatic; pro-14.99-undercuts-chatgpt-plus-claude-pro-by-5-bundles-vibe-coding; student-5.99-verified-edu-first-time-12mo; team-24.99-seat-50-mo-floor-2-seats-min; pro-sub-not-include-api-credits-two-separate-bills; medium-3.5-256k-half-claude-gemini-1m; eu-gdpr-jurisdiction-open-weights-self-host-exit"

4. Access Type

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 Electron app (CLI covers it)

5. Reviews (2026)

Praise: EU/GDPR jurisdiction is the moat​ (Accenture+SAP sovereign-AI deals), $14.99 Pro bundles Vibe coding (ChatGPT charges $20 + Codex separate), Small 4/Medium 3.5 open weights = audit + self-host, multilingual EU langs beat US peers on register, Flux image gen in free tier, Work Mode approval-gated = compliant agentic, 256K ctx covers most doc work, student $5.99 steep discount.

Gripes: free ~25 msg/day soft cap dies mid-session, Pro sub ≠ API credit (two bills surprise devs), Medium 3.5 256K = half Claude/Gemini 1M on massive corpora, no GPTs/Operator/Sora-class voice-video ecosystem, peak reasoning/math trails GPT-5.6/Claude Opus 4.8, mobile app less polished than ChatGPT/Claude, student plan first-time-only 12-mo, Team $50 floor wastes solo-2-seat buyers, third-party connector ecosystem thinner than Zapier/OpenAI.

6. Best for / Not for

7. Competitors