Llama 3.3 70B

Llama 3.3 70B Versatile

Provider: Groq (Meta) Context Window: 128,000 tokens Quality: High Price: $0.59 per 1M input tokens

Overview

Llama 3.3 70B is Meta’s most capable open-source model, running on Groq’s optimized inference. It offers near-proprietary quality at open-source pricing.

Best For

• Complex reasoning and analysis
• Code generation and debugging
• Open-source preferred workloads
• Cost-effective high-quality tasks

Pricing

Tier	Input	Output
BYOK	$0.59/1M	$0.79/1M
Marketplace	$0.6077/1M	$0.8137/1M

Routing Recommendation

• Accurate mode: Strong choice for complex tasks at lower cost than GPT-4o
• Balanced mode: Excellent value proposition
• Cheap mode: Good quality at low cost
• Eco mode: Moderate selection

Why Groq?

Groq runs Llama on custom LPU hardware, delivering:

• Fastest inference speeds for open models
• No GPU required on your end
• Pay-per-token pricing

Usage

curl http://127.0.0.1:8787/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "fastlane-auto", "messages": [{"role": "user", "content": "Write a Python web scraper"}]}'