Mistral AI (La Plateforme)

Mistral offers high-performance European-made models with a developer friendly free tier for prototyping and personal use.

🚀 Key Specs

  • Model Support: Mistral-Large, Mistral-Next, Mixtral 8x22B, Pixtral (Multimodal).
  • Rate Limit (Free Tier):
    • 1 Request Per Second (RPS) overall limit.
    • ~500k-1M Tokens Per Minute window.
  • Trial Allowance: Free allowance for experimentation (no CC needed).

🗝 How to get an API Key

  1. Sign up at Mistral AI (console.mistral.ai).
  2. Go to API Keys and click Create New Key.
  3. Note: If you don’t add billing, your account will remain on the “Free Tier” limits.

🛠 Usage Example (Mistral SDK)

from mistralai.client import MistralClient
from mistralai.models.chat_completion import ChatMessage
 
api_key = "YOUR_MISTRAL_API_KEY"
client = MistralClient(api_key=api_key)
 
messages = [ChatMessage(role="user", content="What is Mistral?")]
chat_response = client.chat(model="mistral-large-latest", messages=messages)
 
print(chat_response.choices[0].message.content)