Llama 3.3 70B

High-performance multilingual language model for chat and reasoning

Model Details

A powerful multilingual language model optimized for dialogue use cases. Outperforms many available open source and closed chat models on common industry benchmarks. Supports common languages with strong performance.

Parameters

70.6 billion

Context Window

64k tokens

Recommended Use

Ideal for assistant-style chat, multilingual tasks, and general text generation. Should be deployed as part of a system with additional safety guardrails.

Supported Languages

English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai

Usage Examples

Installation:

pip install tinfoil

Inference:

from tinfoil import TinfoilAI

client = TinfoilAI(
    enclave="llama3-3-70b-p.model.tinfoil.sh",
    repo="tinfoilsh/confidential-llama3-3-70b-prod",
    api_key="YOUR_API_KEY",
)

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Hello!",
        }
    ],
    model="llama3-3-70b",
)
print(chat_completion.choices[0].message.content)

Tinfoil