Compare AI models

Specifications
Parameters
47B6B
Context window
32k32k
Benchmarks
HumanEval
40.2%9.8%
HellaSwag
84.4%57%
MMLU
70.6%45.2%
GSM8K
58.4%25.9%
Overview
CompanyMistralZ.ai
Release dateDec 8 2023Jun 25 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mixtral 8×7B leads ChatGLM2-6B on 4 of the 4 benchmarks they both report (HumanEval, HellaSwag, MMLU, GSM8K). ChatGLM2-6B shipped 166 days before Mixtral 8×7B, so benchmark comparisons should account for the intervening progress.

Mixtral 8×7B has 47B parameters, while ChatGLM2-6B has 6B. Context windows are 32k (Mixtral 8×7B) vs 32k (ChatGLM2-6B).

On HumanEval, Mixtral 8×7B leads at 40.2% vs ChatGLM2-6B at 9.8%. On HellaSwag, Mixtral 8×7B leads at 84.4% vs ChatGLM2-6B at 57%. On MMLU, Mixtral 8×7B leads at 70.6% vs ChatGLM2-6B at 45.2%. On GSM8K, Mixtral 8×7B leads at 58.4% vs ChatGLM2-6B at 25.9%.