Support Matrix

Hardware

Unless specified otherwise, NVIDIA NIM for vision language models (VLMs) should, but are not guaranteed to, run on any NVIDIA GPU, provided the GPU has sufficient memory. They can also run on multiple homogeneous NVIDIA GPUs with sufficient aggregate memory and a CUDA compute capability of >= 7.0 (8.0 for bfloat16) unless otherwise specified. For more information, refer to Supported Models.

NVIDIA NIM for VLMs does not support NVIDIA Virtual GPU (vGPU) environments.

For information on the supported operating systems, drivers, and software, refer to the About Get Started page.

Supported Models

Gemma-4-26B-A4B-IT

Latest supported release tag: 2.0.11-variant

The following section lists the supported configurations for google/gemma-4-26b-a4b-it (NGC catalog page).

Generic Configuration

The GPU Memory column is per-GPU memory in GB; the Disk Space column is the NGC artifact size needed in the NIM cache (one-time download on first launch), in GB.

GPU	GPU Memory	Precision	# of GPUs	Disk Space
NVIDIA A100 SXM4 40GB	40	BF16	4	52
NVIDIA A100 SXM4 80GB	80	BF16	2	52
NVIDIA A10G	24	BF16	4	52
NVIDIA B200	192	BF16	1	52
NVIDIA B200	192	NVFP4	1	19
NVIDIA B300 SXM6 AC	288	BF16	1	52
NVIDIA B300 SXM6 AC	288	NVFP4	1	19
NVIDIA GB200	192	BF16	1	52
NVIDIA GB200	192	NVFP4	1	19
NVIDIA GB300	288	BF16	1	52
NVIDIA GB300	288	NVFP4	1	19
NVIDIA GH200 144G HBM3e	144	BF16	1	52
NVIDIA GH200 480GB	96	BF16	1	52
NVIDIA H100 80GB HBM3	80	BF16	2	52
NVIDIA H100 NVL	94	BF16	1	52
NVIDIA H200	141	BF16	1	52
NVIDIA H200 NVL	141	BF16	1	52
NVIDIA L40S	48	BF16	4	52
NVIDIA RTX PRO 4500 Blackwell Server Edition	32	BF16	4	52
NVIDIA RTX PRO 4500 Blackwell Server Edition	32	NVFP4	1	19
NVIDIA RTX PRO 6000 Blackwell Server Edition	96	BF16	1	52
NVIDIA RTX PRO 6000 Blackwell Server Edition	96	NVFP4	1	19