Mistral AI Models
Mistral AI produces highly efficient models that punch above their weight class.
Model Lineup
Open Models
| Model | Parameters | Context | Architecture |
|---|---|---|---|
| Mistral 7B | 7B | 32K | Dense transformer |
| Mixtral 8x7B | 47B (12B active) | 32K | MoE |
| Mixtral 8x22B | 141B (39B active) | 64K | MoE |
API-Only Models
| Model | Best For |
|---|---|
| Mistral Large | Complex reasoning |
| Mistral Medium | Balanced performance |
| Mistral Small | Fast, cost-effective |
Key Innovations
Mixture of Experts (MoE)
- Uses 8 expert networks
- Only 2 experts active per token
- Near 8x7B quality at 7B inference cost
Sliding Window Attention
- Efficient long-context handling
- Reduced memory usage
- Better than naive attention
Running Locally
With Ollama
# Standard Mistral
ollama pull mistral
# Mixtral
ollama pull mixtral:8x7b
# Run
ollama run mistralPython Integration
import ollama
response = ollama.chat(
model='mistral',
messages=[
{'role': 'user', 'content': 'Write a haiku about coding'}
]
)
print(response['message']['content'])Mistral API
Setup
from mistralai.client import MistralClient
from mistralai.models.chat_completion import ChatMessage
client = MistralClient(api_key="your-api-key")
response = client.chat(
model="mistral-medium",
messages=[ChatMessage(role="user", content="Hello!")]
)
print(response.choices[0].message.content)Function Calling
tools = [{
"type": "function",
"function": {
"name": "get_weather",
"parameters": {
"type": "object",
"properties": {
"location": {"type": "string"}
},
"required": ["location"]
}
}
}]
response = client.chat(
model="mistral-large-latest",
messages=[ChatMessage(role="user", content="Weather in Paris?")],
tools=tools
)Comparison with Competitors
| Aspect | Mistral 7B | Llama 2 7B | GPT-3.5 |
|---|---|---|---|
| Code | Strong | Good | Strong |
| Reasoning | Good | Good | Strong |
| Speed | Fast | Fast | Medium |
| Cost | Free/Cheap | Free | Paid |
Best Practices
- Use MoE for quality - Mixtral gives excellent results
- Leverage context - 32K window is generous
- Prompt engineering - Works well with chain-of-thought
- Temperature - Default 0.7 works well
intermediateLLM ComparisonUpdated 2024-12-18
Related Guides
Llama Models Guide
Complete guide to Meta's Llama family of open-source language models.
Running LLMs Locally
Set up and run open-source LLMs on your own hardware with Ollama, llama.cpp, and more.
Claude: Capabilities & Best Practices
Master Claude's unique strengths, context window, and optimal use cases.
- mistral
- mistral ai
- mixtral
- open source llm
- moe