Llama 3.3 70B Instruct vs Nova Lite

Llama 3.3 70B Instruct

Llama 3.3 70B Instruct, created by Meta, is a multilingual large language model specifically fine-tuned for instruction-based tasks and optimized for conversational applications. It is capable of processing and generating text in multiple languages, with a context window supporting up to 128,000 tokens. Launched on December 6, 2024, the model surpasses numerous open-source and proprietary chat models in various industry benchmarks. It utilizes Grouped-Query Attention (GQA) to improve scalability and has been trained on a diverse dataset comprising over 15 trillion tokens from publicly available sources. The model's knowledge is current up to December 2023.

Nova Lite

Amazon Nova Lite is a versatile multimodal model designed to process text, image, and video inputs, producing text-based outputs. Featuring a 300K-token context window, it is well-suited for real-time interactions, document analysis, and visual question answering. As part of the Amazon Nova foundation models, it supports fine-tuning and distillation, enabling advanced customization.

	Llama 3.3 70B Instruct	Nova Lite
Provider	Meta	Amazon
Web Site	Open	-
Release Date	Dec 06, 2024 4 months ago	Dec 02, 2024 4 months ago
Modalities	text ?	text ? images ? video ?
API Providers	Fireworks, Together, DeepInfra, Hyperbolic	Amazon Bedrock
Knowledge Cut-off Date	12.2024	Purposefully not disclosed
Open Source	Yes	No
Pricing Input	$0.23 per million tokens	$0.06 per million tokens
Pricing Output	$0.40 per million tokens	$0.24 per million tokens
MMLU	86% 0-shot, CoT Source	80.5% CoT Source
MMLU Pro	68.9% 5-shot, CoT Source	Not available
MMMU	Not available	Not available
HellaSwag	Not available	Not available
HumanEval	88.4% pass@1 Source	85.4% pass@1 Source
MATH	77% 0-shot, CoT Source	73.3% CoT Source
GPQA	50.5% 0-shot, CoT Source	42% Main Source
IFEval	92.1% Source	89.7% Source
Mobile Application	-	-

Llama 3.3 70B Instruct

Nova Lite

Provider

Compare LLMs
Llama 3.3 70B Instruct vs Nova Lite