Google Gemini is a family of multimodal large language models developed by Google DeepMind, serving as the successor to LaMDA and PaLM 2. Gemini Pro was announced on December 13, 2023, positioned as a competitor to OpenAI's GPT-4.
vs.
The Llama 3 Instruct 8B model, developed by Meta, features an 8K tokens input context window and comprises 8 billion parameters. Officially launched on April 18th, 2024, this model operates under an open license, enabling broad accessibility and utilization across diverse applications and projects.
Overview
Overview of the AI models
Gemini Pro
Llama 3 Instruct 8B
Release Date
When the model was released.
December 13th, 2023
April 18th, 2024
Input Context Window
The quantity of tokens that the input context window can accommodate.
33K
tokens
8K
tokens
License
Terms and conditions under which an AI model can be used.
Proprietary
Open
Benchmark
Compare important benchmarks among the AI models
Gemini Pro
Llama 3 Instruct 8B
MMLU
Measuring Massive Multitask Language Understanding (MMLU) is a benchmark for evaluating the capabilities of language models.
72
68
Latency
Seconds taken to receive the first tokens, measured on an input size of 1000 tokens.
