Mistral AI

• June 24, 2026: Fresh pricing-page check found no material change versus the June 23 buyer frame: Mistral still lists Free, Pro $14.99/month, Education $5.99/month, Team $24.99/user/month, Enterprise custom, Studio/API examples, model cards, and pricing tables can drift.
• June 23, 2026: Fresh source check confirmed the Mistral buyer frame remains unchanged: Vibe Free/Pro/Education/Team, Studio/API pricing, Search Toolkit, Vibe for Code, Large 3, Medium 3.5, Small 4, and the pricing-page versus model-card drift buyers should verify in Studio before budgeting.
• June 21, 2026: Rechecked Mistral pricing, Vibe, model docs, Search Toolkit, and legal/data-posture language during the comparison-policy cleanup. Mistral remains the native model-provider/platform lane, while Poe remains the multi-model chat and bot aggregator lane, so that pair is blocked as a direct comparison page. This refresh also tightens old GDPR Article 44 wording and clarifies that Voxtral TTS commercial and internal-business use should still get license review because the default model card uses CC BY-NC 4.0.
• June 15, 2026: Current official sources still consolidate Le Chat into Mistral Vibe, and the Vibe agent announcement says Chat mode will soon be sunset while Le Chat history is preserved in Work Mode. The pricing page still lists Free, Pro at $14.99/month, Education at $5.99/month for eligible verified students, Team at $24.99/user/month, Enterprise custom, PAYG credits past plan limits, and per-model API rates. It also lists Large 3 at $0.50/M input and $1.50/M output, Medium 3.5 at $1.50/M input and $7.50/M output, and Small 4 at $0.10/M input and $0.30/M output; the Small 4 model card still shows $0.15/M and $0.60/M, and the pricing FAQ still uses a generic Mistral Large $2/M and $6/M example, so budget from the live Studio quote before committing.
• June 15, 2026: Added Vibe Work/Code details from the May 28 Vibe agent post: Work Mode, Code Mode, VS Code extension, CLI updates, Skills, custom modes/subagents, editable plans, session-scoped permissions, and /teleport. The AI Now Summit post also adds a procurement signal: Mistral says the Les Ulis inference facility is a 10 MW site scheduled for Q3 2026.
• June 15, 2026: Reconfirmed the Mistral 3 timeline. The official Mistral 3 launch post is dated December 2, 2025, while April 28, 2026 is the Mistral Medium 3.5 model-card date.
• June 6, 2026: Mistral pricing, model docs, Mistral 3, and Search Toolkit were rechecked while refreshing Mistral AI vs Qwen. The buyer split is now explicit: Mistral is the EU/open-model/vendor-platform lane; Qwen is the Alibaba/Qwen Cloud, multilingual, and model-family lane.
• June 5, 2026: Mistral pricing and docs were rechecked while refreshing DeepSeek vs Mistral AI. The buyer split is now explicit: DeepSeek experiment; Mistral is the more procurement-friendly EU model platform with Vibe, Studio, Search, open weights, and enterprise paths.
• June 3, 2026: Mistral pricing recheck kept Pro at $14.99/month and Team at $24.99/user/month, while the public page now explicitly surfaces Mistral Vibe, Vibe for Code, Studio, Search, and API pricing beside Le Chat.
• May 28, 2026: Mistral launched Search Toolkit, a hosted search/RAG layer for semantic search, vector search, and web-search-enhanced applications.
• May 13, 2026: Le Chat/Vibe consumer pricing migrated to USD and added a Team tier at $24.99/mo per seat (30GB storage per user, domain verification, data export). Pro held at $14.99/mo; Enterprise remains custom.
• May 12, 2026: Mistral published an SDK supply-chain advisory tied to the TanStack incident. Mistral says its infrastructure was not compromised, but compromised npm and PyPI package versions were briefly published. Teams using Mistral SDKs should check lockfiles, images, caches, and deployed artifacts for affected versions and rotate exposed secrets where needed.
• December 2, 2025: Mistral 3 launched with Mistral Large 3 and new Ministral edge models. The Apache 2.0 release made Large 3 a 41B-active / 675B-total MoE flagship, added Ministral 3 at 3B/8B/14B with base, instruct, and reasoning variants, and went live across Mistral AI Studio, Amazon Bedrock, Azure Foundry, Hugging Face, OpenRouter, Fireworks, Together AI, Modal, IBM watsonx, and Unsloth AI on day one.

Pick Mistral if EU data sovereignty is non-negotiable or you want frontier-tier open weights. under a Modified MIT license, and Small 4/Ministral 3 keep the cheaper multimodal and edge lanes alive. That removes a lot of vendor-lock pressure for regulated teams, but production buyers still need model-by-model license review.

Skip it for the deepest everyday-assistant ecosystem. ChatGPT remains the easier broad assistant. Gemini wins Google Workspace. Claude wins careful long-form writing. Mistral’s advantage is model control, European posture, Vibe for Code, Search Toolkit, Studio/API access, and open-weight strategy.

Who pays which tier: Free for casual Vibe chat/search/creation/coding trials, Pro $14.99/mo for individuals, Education $5.99/mo for verified students, Team $24.99/user/mo for shared workspaces, Mistral Studio/API/Search Toolkit for developers, and Enterprise/Forge for procurement, deployment, and custom model work.

Every volatile model, pricing, and platform data point verified on 2026-06-24 against mistral.ai/pricing, Mistral Vibe, the Vibe agent announcement, AI Now Summit 2026, docs.mistral.ai/models/overview, the Large 3 / Medium 3.5 / Small 4 model cards, Mistral 3, Mistral Search Toolkit, and Mistral security advisories.

• Hosted prices vary by platform and source. Large 3, Medium 3.5, Small 4, and Ministral 3 are open-weight, so per-token rates depend on whether you use Mistral AI Studio, Bedrock, Azure Foundry, Fireworks, Together AI, OpenRouter, or self-host. Benchmark before committing.
• Small 4 has a live source discrepancy. On June 23, 2026 the Mistral pricing page listed $0.10/M input and $0.30/M output, while the Small 4 model card still listed $0.15/M and $0.60/M. Treat Studio billing as authoritative for purchase math.
• Large-family pricing examples can mislead. The pricing FAQ still gives a generic Mistral Large example at $2/M input and $6/M output, while the Large 3 model card and API table list $0.50/M and $1.50/M. Budget by exact model ID and live quote, not by prose examples.
• Voxtral license is CC BY-NC 4.0. Non-commercial only. Using it in a commercial product (even internally) requires a commercial license from Mistral.
• Smaller plugin ecosystem. la Plateforme is OpenAI-compatible, so most tools work, but Mistral-first integrations are thin compared to ChatGPT or Claude.
• Vibe still trails the broadest consumer ecosystems. Functional and now much broader than old Le Chat, but ChatGPT, Claude.ai, and Gemini still have deeper everyday-assistant ecosystems. GPT Store-style distribution remains limited.
• Lineup churn is real. Medium 3.1 (v25.08), Medium 3.5 (v26.04), Devstral aliases, and Large 3/Small 4 overlap for adjacent workloads, and the model overview lists retirement dates for some exact IDs. Read the current model docs before locking in a model ID for a long-running deployment.
• Chat mode is not the long-term frame. Mistral’s Vibe agent post says Chat mode will soon be sunset, with history preserved in Work Mode. Buyers should evaluate Vibe Work/Code, not an old Le Chat mental model.
• Smaller capital base than U.S. frontier labs. ~$1B raised vs tens of billions at OpenAI, Anthropic, and Google DeepMind. Compute ceiling is a real constraint on frontier model cadence.
• No native video generation. For multimodal output including video, workflow needs a second tool.

This page was produced by the aipedia.wiki editorial pipeline, an automated system that ingests vendor documentation, verifies pricing and model details against primary sources, and generates the editorial analysis you are reading. No individual human wrote this review. Scoring follows the four-dimension rubric at /about/scoring/ (Utility, Value, Moat, Longevity, unweighted average). Last verified 2026-06-24 against mistral.ai/pricing, Mistral Vibe, the Vibe agent announcement, AI Now Summit 2026, docs.mistral.ai/models/overview, current model cards, Mistral Search Toolkit, the Mistral 3 launch post, and Mistral security advisories.

Is Mistral free? Yes, in two ways. Vibe has a Free plan for limited chat, web search, creation, and terminal coding. Mistral also publishes open-weight models including Large 3, Medium 3.5, Small 4, Ministral 3, Devstral, Voxtral, and older Mistral/Mixtral families, but commercial self-hosting terms should be checked per model.

What is Mistral 3? The December 2, 2025 open-model release. Mistral Large 3 is the flagship: a 41B-active / 675B-total mixture-of-experts model under Apache 2.0, the first MoE from Mistral since the Mixtral line. The Ministral 3 family ships alongside at 3B, 8B, and 14B parameter sizes with base, instruct, and reasoning variants, image understanding, and multilingual support. All were available across Mistral AI Studio, Amazon Bedrock, Azure Foundry, Hugging Face, OpenRouter, Fireworks, Together AI, Modal, IBM watsonx, and Unsloth AI on day one.

What is Mistral Small 4? A single Apache 2.0 model released March 16, 2026 that unifies instruct, reasoning, and coding capabilities. The model card lists 119B parameters with 6.5B active and 256K context. The June 15 pricing page lists $0.10/M input and $0.30/M output, while the Small 4 model card still lists $0.15/M and $0.60/M.

What about Mistral Medium 3.5? v26.04 frontier-class multimodal model optimized for agentic and coding workloads. It is a 128B dense model, released April 28, 2026 as open weights under a Modified MIT license, with 256K context and $1.50/M input plus $7.50/M output on Mistral’s pricing page.

How does Mistral compare to OpenAI frontier models? Mistral Large 3 is broadly competitive on text benchmarks and now ships under Apache 2.0. OpenAI frontier models win on ecosystem, image generation, and multimodal breadth. For EU-regulated workloads, Mistral is the default non-U.S. provider to evaluate because it combines EU incorporation, private deployment options, DPA/SCC language, and open-weight flagship options.

Why choose Mistral over OpenAI or Anthropic? Three concrete reasons: open weights at the frontier, EU-oriented deployment and data safeguards, and a credible edge model lineup via Ministral 3. U.S. developers without compliance constraints should weigh those against OpenAI’s ecosystem or Anthropic’s reasoning lead.

What is Forge? Mistral’s enterprise fine-tuning platform launched March 17, 2026. Enables custom-model training on proprietary data under a license-fee model. Contact Mistral sales for specifics.

Is Voxtral safe for commercial products? Not under the default license. CC BY-NC 4.0 restricts commercial use. Internal business tools can still be commercial use, so review the exact license and get a commercial license from Mistral when needed.

• Mistral AI pricing (verified 2026-06-24): Free, Pro, Education, Team, Enterprise, Vibe, Vibe for Code, Studio, Search, and API pricing
• Mistral Vibe (verified 2026-06-24): Vibe formerly Le Chat, Vibe for Code, connectors, MCP, remote agents, and enterprise deployment notes
• Mistral Vibe agent announcement (verified 2026-06-24): Work Mode, Code Mode, VS Code extension, CLI updates, Skills, session permissions, /teleport, and Chat-mode sunset language
• Mistral AI Now Summit 2026 (verified 2026-06-24): Vibe unification, industrial AI context, and Les Ulis inference-facility timing
• Mistral model overview (verified 2026-06-24): current production lineup plus legacy/deprecation rows
• Mistral Medium 3.5 model card (verified 2026-06-24): 128B dense model, Modified MIT, context, and per-token card pricing
• Mistral Small 4 model card (verified 2026-06-24): context, parameter count, and model-card pricing
• Mistral Large 3 model card (verified 2026-06-24): launch date, context, parameter count, and per-token card pricing
• Mistral 3 launch post (verified 2026-06-24): Large 3 and Ministral 3 announcement
• Mistral Search Toolkit (verified 2026-06-24): semantic search, vector search, RAG, and web search announcement
• Mistral security advisories (verified 2026-06-24): SDK supply-chain advisory status and affected package versions
• Hugging Face · mistralai (verified 2026-06-24): open-weight model releases

• Category: AI Chatbots
• Comparisons: DeepSeek vs Mistral AI · Mistral AI vs Qwen