Mistral AI Pricing Guide: Le Chat Pro, API Token Rates, and Open Weights (2026)
Headquartered in Paris, Mistral AI is the pioneer of high-performance European artificial intelligence. By combining openly downloadable model weights with commercial enterprise APIs on La Plateforme, Mistral provides engineering teams with complete deployment sovereignty.
Whether choosing between the consumer Le Chat Pro subscription or deploying Codestral and Mistral Large 2 via managed APIs, this guide breaks down all pricing structures and cost optimization pathways.
1. Mistral Le Chat Consumer Plans
┌────────────────────────────────────────────────────────────────────────┐
│ MISTRAL LE CHAT ACCESS TIERS │
└───────────────────────────────────┬────────────────────────────────────┘
│
┌─────────────────────────┴─────────────────────────┐
▼ ▼
┌──────────────────────────────┐ ┌──────────────────────────────┐
│ LE CHAT FREE │ │ LE CHAT PRO ($9.99/MO) │
├──────────────────────────────┤ ├──────────────────────────────┤
│ • Mistral Large & Small │ │ • Priority High-Speed Queue │
│ • Live Web Search │ │ • Unlimited Canvas Workspace │
│ • Document Analysis │ │ • Flux Image Generation │
└──────────────────────────────┘ └──────────────────────────────┘| Plan | Monthly Cost | Model Access | Key Features | Best For |
|---|---|---|---|---|
| Le Chat Free | $0.00 | Mistral Large, Mistral Small | Web search, basic Canvas, doc analysis | Daily questions & testing |
| Le Chat Pro | $9.99 / mo | Mistral Large 2, Codestral | Zero queue delays, high-res Flux images | Technical power users & writers |
2. La Plateforme API Token Pricing Matrix
For production backends, IDE plugins, and agent systems, Mistral charges per 1 million tokens processed:
| Model | Focus / Capability | Context Window | Input Tokens / 1M | Output Tokens / 1M |
|---|---|---|---|---|
| Mistral Large 2 | Frontier Reasoning & Multilingual | 128,000 | $2.00 | $6.00 |
| Codestral 2501 | Specialized Software Engineering | 256,000 | $0.30 | $0.90 |
| Mistral Small 3 | Fast High-Throughput Routing | 32,000 | $0.20 | $0.60 |
| Pixtral Large | High-Res Vision & OCR | 128,000 | $2.00 | $6.00 |
| Mistral Embed | Semantic Text Embeddings | 8,000 | $0.10 | N/A |
[!TIP] 50% Batch API Discount: Submitting background jobs via the Mistral Batch API cuts token rates in half across all models (e.g. Codestral drops to $0.15 / 1M input and $0.45 / 1M output).
3. Commercial API vs Open-Weight Self-Hosting
One of Mistral’s greatest strengths is licensing flexibility:
- Open Weights (Apache 2.0): Models such as Mistral NeMo 12B and Mistral 7B are 100% open-source under the Apache 2.0 license, allowing free commercial deployment on cloud providers like RunPod.
- Mistral Research License (MRL): Available for free non-commercial research, or accessible commercially via La Plateforme API or Microsoft Azure / AWS Bedrock endpoints.
