Back to list
modelDeepSeekMoEopen sourcehigh performance
DeepSeek-V3
In-depth exploration of open source MoE large models, 671B parameters and 37B activations, performance comparable to GPT-4
21 views0 stars3/5/2026
In-depth exploration of open source MoE large models, 671B parameters and 37B activations, performance comparable to GPT-4