How much memory do open models need?

Pick a model to see every rig that runs it, with estimated speed and current prices.

ModelTotal / active params~Memory at 4-bit~Memory at 8-bitRigs that fit (4-bit)
Qwen3.5 9BAlibaba Qwen9B / 9B7 GB12 GB18 of 18
Gemma 4 12BGoogle DeepMind12B / 12B9 GB15 GB18 of 18
GPT-OSS 20BOpenAI21B / 3.6B13 GB25 GB18 of 18
Gemma 4 26B-A4BGoogle DeepMind25.2B / 3.8B18 GB30 GB17 of 18
Qwen3.8 27BAlibaba Qwen27B / 27B19 GB32 GB17 of 18
GLM-4.7-Flash 30B-A3BZhipu (Z.ai)30B / 3B21 GB35 GB17 of 18
Gemma 4 31BGoogle DeepMind31B / 31B21 GB36 GB17 of 18
Qwen3.6 35B-A3BAlibaba Qwen35B / 3B24 GB41 GB16 of 18
Llama 3.3 70BMeta70B / 70B46 GB80 GB13 of 18
Qwen3-Next 80B-A3BAlibaba Qwen80B / 3B52 GB91 GB12 of 18
Llama 4 Scout 109B-A17BMeta109B / 17B71 GB123 GB12 of 18
GPT-OSS 120BOpenAI117B / 5.1B67 GB132 GB12 of 18
Nemotron 3 Super 120B-A12BNVIDIA120B / 12B78 GB135 GB11 of 18
Qwen3.5 122B-A10BAlibaba Qwen122B / 10B79 GB138 GB11 of 18
Qwen3 235B-A22BAlibaba Qwen235B / 22B151 GB264 GB1 of 18
Muse Spark Open 816B-A42BMeta816B / 42B521 GB912 GB0 of 18

Restock & price-drop alerts

One short email a week: what got cheaper, what's back in stock, and which new open models your machine can run. Unsubscribe anytime.