Mixtral 8x22B v0.1

Find hardware that runs Mixtral 8x22B v0.1 on your budget

Reference: Model card on Hugging Face · Published benchmark scores

Parameters: 140.62 B
Active per token: 39.15 B
Architecture: MoE
Modality: text
Max context: 65536
Released: 2024-04-16
License: apache-2.0

Measurements

Decode is token generation — the speed you feel while an answer streams. Prefill is prompt processing — the wait before it starts. Why bandwidth predicts decode speed.

What verified, single-source and estimated mean, and the same rows with every filter and sort in the benchmarks explorer.

Speed over time

Cheapest hardware by target speed

The lowest-priced entry that reaches each threshold, using the strongest evidence available. Estimated rows are marked.