Compare Llama 3.1 70B Instruct vs. Chat Bison

Pricing About Us Blog Docs

Compare

Understand and compare Llama 3.1 70B Instruct vs. Chat Bison

Interested in comparing the performance of these models in production? Get a demo of Context.ai

Overview

Llama 3.1 70B Instruct was released about 1 year after Chat Bison.

	Llama 3.1 70B Instruct	Chat Bison
Provider The entity that provides this model.	Meta	Google
Input Context Window The number of tokens supported by the input context window.	128K tokens	8,192 characters
Maximum Output Tokens The number of tokens that can be generated by the model in a single request.	2,048 tokens	2,048 characters
Release Date When the model was first released.	July 23rd, 2024 about 2 months ago 2024-07-23	July 10th, 2023 about 1 year ago 2023-07-10
Knowledge Cutoff Limit on the knowledge base used by the model.	December 2023 December 2023	Mid 2021 Mid 2021
Open Source	Yes	No
API Providers The providers that offer this model. (This is not an exhaustive list.)	Azure AI, AWS Bedrock, Google Cloud Vertex AI Model Garden, NVIDIA NIM, IBM watsonx	Google Cloud Vertex AI Model Garden

Pricing

	Llama 3.1 70B Instruct	Chat Bison
Input Cost of input data provided to the model.	Pricing not available.	Pricing not available.
Output Cost of output tokens generated by the model.	Pricing not available.	Pricing not available.

Benchmarks

Compare relevant benchmarks between Llama 3.1 70B Instruct and Chat Bison.

	Llama 3.1 70B Instruct	Chat Bison
MMLU Evaluating LLM knowledge acquisition in zero-shot and few-shot settings.	83.6 (5-shot) Source	78.3 (5-shot) Source
MMMU A wide ranging multi-discipline and multimodal benchmark.	Benchmark not available.	Benchmark not available.
HellaSwag A challenging sentence completion benchmark.	Benchmark not available.	Benchmark not available.
GSM8K Grade-school math problems benchmark.	95.1 (8-shot) Source	Benchmark not available.
HumanEval A benchmark to measure functional correctness for synthesizing programs from docstrings.	Benchmark not available.	Benchmark not available.
MATH Benchmark performance on Math problems ranging across 5 levels of difficulty and 7 sub-disciplines.	68.0 (0-shot) Source	Benchmark not available.