SOURCE-REVIEWED MODEL PROFILE
Mixtral-8x7B-Instruct-v0.1
An instruction-tuned mixture-of-experts text model that routes each token through selected experts.
Metadata reviewed 2026-10-24 · Original creator: mistralai
What it does
- Sparse mixture-of-experts
- Text chat and coding prompts
- Instruction following
Local fit
Large GPU / more RAM. Try a multi-character story workshop. Sparse activation reduces computation but does not remove the full expert weight-storage requirement.
Configuration value; backend support and enough memory are required. Start with a smaller context on limited hardware.
Evidence and limits
Source reviewed; no local generation benchmark for this exact checkpoint.
Connect a supported local text/vision backend; not independently tested
License and source
Keep upstream notices and check the exact downloaded artifact and parent-model terms. The original creator remains credited.
Read the original model card ↗
Reviewed source revision: eba92302a2861cdc0098cc54bc9f17cb2c47eb61