← Back to Directory
Meta AI
LLM📅 Released: 2024-12-06

Llama 3.3 70B

Llama 3.3 70B provides Llama 3.1 405B level performance in a smaller package.

#open-source#efficient#popular

Overview

A refined version of the 70B model.

Unique Factor

Key features include: Distilled intelligence, 128K context, Open weight

Key Capabilities

●
Distilled intelligence
●
128K context
●
Open weight

Benchmarks

MMLU Score
88%
HumanEval (Coding)
88%
GPQA Diamond
74%
MATH Benchmark
82%

Top Use Cases

General Purpose

Used for various AI tasks related to LLM

Example: “Tell me more about Llama 3.3 70B”

Detailed Features

01

Distilled intelligence

02

128K context

03

Open weight

✓ Strengths & Pros

  • • High performance
  • • Reliable architecture

✕ Limitations & Cons

  • • Standard limitations

Ideal Usage & Target Audience

Best For

Researchers and developers.

Not Recommended For

Users looking for high-latency solutions.

API Implementation

text
# API access for Llama 3.3 70B

Check the official documentation for full SDK details.

Quick Links

Technical Specs

Context128,000 tokens
Params70B
LicenseLlama 3.3 Community
ArchTransformer

API Pricing

$0 / 1M input tokens

Output: $0 / 1M tokens

✓ Free tier available
Access API

Developer

The open-source champions — creators of Llama, the world's most popular open-weights model family.

Prompt Library

Browse Coding Prompts →

📋

Previous Version

Llama 3 1 70b →