Understand and compare
Mistral 8x7B Instruct
vs.
GPT-4 Turbo 2024-04-09
Try
Podial
Turn your documents into engaging podcast discussions.
Overview
Mistral 8x7B Instruct was released
4 months before
GPT-4 Turbo 2024-04-09.
Mistral 8x7B Instruct
|
GPT-4 Turbo 2024-04-09
|
|
---|---|---|
Provider
The entity that provides this model.
|
Mistral
|
OpenAI
|
Input Context Window
The number of tokens supported by the input context window.
|
32K
tokens
|
128K
tokens
|
Maximum Output Tokens
The number of tokens that can be generated by the model in a single request.
|
4,096
tokens
|
4,096
tokens
|
Release Date
When the model was first released.
|
2023-12-11
|
2024-04-09
|
Knowledge Cutoff
Limit on the knowledge base used by the model.
|
Unknown
|
December 2023
|
Open Source
|
|
|
API Providers
The providers that offer this model. (This is not an exhaustive list.)
|
|
|
Pricing
Mistral 8x7B Instruct is
roughly 14x cheaper compared
to GPT-4 Turbo 2024-04-09 for input tokens and
roughly 43x cheaper
for output tokens.
Mistral 8x7B Instruct
|
GPT-4 Turbo 2024-04-09
|
|
---|---|---|
Input
Cost of input data provided to the model.
|
$0.70
per million tokens
|
$10.00
per million tokens
|
Output
Cost of output tokens generated by the model.
|
$0.70
per million tokens
|
$30.00
per million tokens
|
Benchmarks
Compare relevant benchmarks between Mistral 8x7B Instruct
and GPT-4 Turbo 2024-04-09.
Mistral 8x7B Instruct
|
GPT-4 Turbo 2024-04-09
|
|
---|---|---|
MMLU
Evaluating LLM knowledge acquisition in zero-shot and few-shot settings.
|
70.6
(5-shot)
|
86.7
(5-shot)
|
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.
|
Benchmark not available.
|
Benchmark not available.
|
HumanEval
A benchmark to measure functional correctness for synthesizing programs from docstrings.
|
Benchmark not available.
|
88.2
(0-shot)
|
MATH
Benchmark performance on Math problems ranging across 5 levels of difficulty and 7 sub-disciplines.
|
Benchmark not available.
|
73.4
(0-shot)
|
Mistral 8x7B Instruct, developed by Mistral, features a context window of 32000 tokens. The model costs 0.07 cents per thousand tokens for both input and output. It was released on December 11, 2023, and achieved a score of 70.6 in the MMLU benchmark in a 5-shot scenario.
GPT-4 Turbo 2024-04-09, developed by OpenAI, features an impressive context window of 128,000 tokens. The model costs 1.0 cent per thousand tokens for input and 3.0 cents per thousand tokens for output. It is set to be released on April 9, 2024.
Measure & Improve LLM
Product Performance.
Get Started