Back to Models
mistralai logo
mistralai/mixtral-8x7b-instruct
Not Available

Mixtral 8x7B Instruct

Mixtral 8x7B Instruct is a pretrained generative Sparse Mixture of Experts, by Mistral AI, for chat and instruction use. Incorporates 8 experts (feed-forward networks) for a total of 47 billion parameters. Instruct model fine-tuned by Mistral. #moe

12/10/2023
32,768 tokens
#245 Text (french)

Specifications

Modalities

Input
text
Output
text

Supported Parameters

frequency_penalty
logit_bias
max_tokens
min_p
presence_penalty
repetition_penalty
response_format
seed
stop
temperature
tool_choice
tools
top_k
top_p

Max Output Tokens

16,384

Frequently asked questions

Is Mixtral 8x7B Instruct available on FastMetal?
Not at the moment. Voxtral Mini 3B, from the same lab, is available on the FastMetal API today.
What is the context window of Mixtral 8x7B Instruct?
32,768 tokens, shared between the prompt and the response.
How does Mixtral 8x7B Instruct rank?
#255 on the public arena's Japanese board (ELO 1,003). Ranks move as the leaderboard is updated.

Leaderboard

Text
OverallELO: 1,197
#333
JapaneseELO: 1,003
#255
ChineseELO: 1,154
#337
KoreanELO: 1,062
#256
EnglishELO: 1,230
#332
frenchELO: 1,239
#245
germanELO: 1,185
#262
spanishELO: 1,174
#264
russianELO: 1,160
#329
CodingELO: 1,239
#334
MathELO: 1,191
#321
Creative WritingELO: 1,159
#332
Instruction FollowingELO: 1,180
#332
Hard PromptsELO: 1,210
#335
Multi-TurnELO: 1,167
#329