Compare AI models

Specifications
Parameters
7B/67B7B
Context window
4k32k
Benchmarks
HumanEval
73.8%38.4%
MATH
32.6%13.2%
Overview
CompanyDeepSeekMistral
Release dateNov 29 2023May 22 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek-LLM leads Mistral 7B v0.3 on 2 of the 2 benchmarks they both report (HumanEval, MATH). DeepSeek-LLM shipped 175 days before Mistral 7B v0.3, so benchmark comparisons should account for the intervening progress.

DeepSeek-LLM has 7B/67B parameters, while Mistral 7B v0.3 has 7B. Context windows are 4k (DeepSeek-LLM) vs 32k (Mistral 7B v0.3).

On HumanEval, DeepSeek-LLM leads at 73.8% vs Mistral 7B v0.3 at 38.4%. On MATH, DeepSeek-LLM leads at 32.6% vs Mistral 7B v0.3 at 13.2%.