Compare models

Pick two or three models to compare license, hardware, and live stats side by side.

 Llama 3.1 8B InstructMistral 7B Instruct v0.3
OrgMetaMistral AI
Params8B7B
ModalityTextText
LicenseConditionalLlama 3.1 Community LicensePermissiveApache 2.0
HardwareConsumer GPUConsumer GPU
Released2024-072024-05
TakeThe default pick for a local 8B chat model: strong instruction following for its size, easy to quantize and run on a single consumer GPU.The classic fully-permissive 7B baseline; not the strongest 7B anymore but has no usage restrictions at all.
Downloads
Likes