TPSTokens Per Second

RakutenAI-3.0: hardware requirements

RakutenAI-3.0 needs about 415 GB of memory at Q4 quantization — a 404 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.

Parameters
671B
Tier
Server only
Licence
Apache-2.0
Released
2026-03-16

Will it run on your machine?

8GB RAMNot enough memory16GB RAMNot enough memory32GB RAMNot enough memory64GB RAMNot enough memory128GB RAMNot enough memory8GB VRAMNot enough memory12GB VRAMNot enough memory16GB VRAMNot enough memory24GB VRAMNot enough memory32GB VRAMNot enough memory

Download sizes by quantization

Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?

QuantDownloadRAM @ 4KRAM @ 32KSource
Q4_K_Mrecommended404 GB415 GB465 GBmeasured
Q5_K_M475 GB486 GB536 GBmeasured
Q6_K551 GB561 GB611 GBmeasured

“Measured” sizes are read from published GGUF files (mmnga-o/RakutenAI-3.0-gguf). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.

Model card on HuggingFace →API specs & pricing →Verified 2026-08-01