Which is better for local AI, the Intel Arc Pro B60 or the NVIDIA GeForce RTX 3090?

Choose the NVIDIA GeForce RTX 3090 for speed. Choose the Intel Arc Pro B60 to spend less.

Intel Arc Pro B60 vs NVIDIA GeForce RTX 3090

For running models locally, priced in Canada. At 64k context the first runs 40 models fully in memory, the second 40.

In short

Both have 24 GB of memory for models; the Intel Arc Pro B60 has 456 GB/s of bandwidth and the NVIDIA GeForce RTX 3090 936 GB/s. At 64k context both run the same 40 models we track fully in memory. On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 2.5× as fast as the Intel Arc Pro B60 (estimated, median across models). In Canada the Intel Arc Pro B60 costs $985 at Amazon and the NVIDIA GeForce RTX 3090 $2,800 at Amazon: the Intel Arc Pro B60 is $1,815 cheaper.

Updated October 15, 2026. Speeds are estimates; how they are worked out.

Intel Arc Pro B60

$985

Amazon

Add to build

NVIDIA GeForce RTX 3090

$2,800

Amazon

Add to build

Specifications

Spec	Intel Arc Pro B60	NVIDIA GeForce RTX 3090
GPU	Arc Pro B60	GeForce RTX 3090
Memory	24 GB GDDR6	24 GB GDDR6X
Memory bus	192-bit	384-bit
Memory bandwidth	456 GB/s	936 GB/s
FP16 compute	98.5 TFLOPS	71 TFLOPS
Software	SYCL / oneAPI	CUDA
Board power	200 W	350 W
PCIe generation	Gen 5	Gen 4
PCIe lanes	x8	x16

What each runs

Estimated decode speed with the most precise quant that fits each one est.

Model	Intel Arc Pro B60	NVIDIA GeForce RTX 3090
Qwen 3.8 27B	10tok/sQ4_K_M	27tok/sQ4_K_M
Qwen 3.6 27B	11tok/sQ4_K_M	24tok/sQ5_K_M
Qwen 3.6 35B-A3B	39tok/sUD-Q4_K_M	107tok/sQ4_K_M
Qwen 3.5 27B	11tok/sQ4_K_M	27tok/sQ4_K_M
Gemma 4 26B-A4B	26tok/sUD-Q5_K_M	72tok/sUD-Q5_K_M
Qwen 3.5 35B-A3B	39tok/sQ4_K_M	107tok/sQ4_K_M
GLM 4.7 Flash	19tok/sQ4_K_M	53tok/sQ4_K_M
gpt-oss 20B	31tok/sMXFP4	86tok/sMXFP4
Gemma 4 12B	16tok/sQ8_0	41tok/sQ8_0
Qwen 3.5 9B	19tok/sQ8_0	50tok/sQ8_0
Gemma 3 27B	10tok/sQ4_K_M	26tok/sQ4_K_M
Ministral 3 14B	10tok/sQ6_K	27tok/sQ6_K
Qwen3 14B	10tok/sQ6_K	25tok/sQ6_K
DeepSeek R1 0528 Qwen3 8B	12tok/sQ8_0	32tok/sQ8_0
Qwen3 8B	12tok/sQ8_0	32tok/sQ8_0
Phi-4 14B	10tok/sQ4_K_M	24tok/sQ5_K_M
Qwen 3.5 4B	34tok/sQ8_0	87tok/sQ8_0
Ministral 3 8B	13tok/sQ8_0	32tok/sQ8_0
Granite 4.2 8B	11tok/sQ8_0	29tok/sQ8_0
Gemma 4 E4B	23tok/sQ8_0	57tok/sQ8_0
LFM2.5 8B-A1B	29tok/sQ8_0	81tok/sQ8_0
Llama 3.1 8B	15tok/sQ6_K	38tok/sQ6_K
Qwen2.5 0.5B	171tok/sQ8_0	434tok/sQ8_0
Qwen3.5 0.8B	139tok/sQ8_0	355tok/sQ8_0
Qwen3.5 2B	101tok/sQ5_K_M	257tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B	60tok/sQ8_0	154tok/sQ8_0
MiniCPM5 2B	42tok/sQ8_0	106tok/sQ8_0
Mellum2.1 12B A2.5B Thinking	34tok/sQ8_0	96tok/sQ8_0
LFM2.5 2.6B	33tok/sQ8_0	84tok/sQ8_0
LFM2.5 VL 3B	33tok/sQ8_0	84tok/sQ8_0
granite vision 4.1 4b	25tok/sQ8_0	65tok/sQ8_0
granite 4.2 3b	25tok/sQ8_0	63tok/sQ8_0
Kimi VL A3B	24tok/smlx-8bit	66tok/smlx-8bit
DeepSeek R1 Distill Qwen 7B	19tok/sQ8_0	49tok/sQ8_0
LFM2.5 24B A2B	19tok/sQ4_K_M	50tok/sQ5_K_M
GLM 4.6V Flash	18tok/sQ8_0	46tok/sQ8_0
GLM 4.1V 9B Thinking	16tok/smlx-8bit	40tok/smlx-8bit
DeepSeek R1 Distill Llama 8B	13tok/sQ8_0	34tok/sQ8_0
Ministral 3 8B Reasoning 2512	13tok/sQ8_0	32tok/sQ8_0
DeepSeek R1 Distill Qwen 14B	10tok/sQ5_K_M	25tok/sQ5_K_M
Qwen 3.5 122B-A10B	offload	offload
Qwen3 Coder Next	offload	offload
Gemma 4 31B	offload	offload
Mistral Small 4 119B	offload	offload
Qwen3 Next 80B-A3B	offload	offload
Devstral Small 2 24B	offload	offload
Nemotron 3.5 Lightning 30B-A3B	offload	offload
Kimi Linear 48B-A3B	offload	offload
Qwen3 Coder 30B-A3B	offload	offload
Qwen3 32B	offload	offload
Mistral Small 3.2 24B	offload	offload
DeepSeek R1 Distill Qwen 32B	offload	offload
Devstral Small 2507	offload	offload
Magistral Small 2506	offload	offload
Magistral Small 2507	offload	offload
Magistral Small 2509	offload	offload
granite 4.2 30b	offload	offload
DeepSeek V4 Flash	won't run	won't run
GLM 5.3 Flash	won't run	won't run
Qwen 3.8 Flash Next	won't run	won't run
MiniMax M2.7	won't run	won't run
Nemotron 3 Super 120B-A12B	won't run	won't run
gpt-oss 120B	won't run	won't run
GLM 4.5 Air	won't run	won't run
Llama 3.3 70B	won't run	won't run
DeepSeek R1	won't run	won't run
DeepSeek R1 0528	won't run	won't run
DeepSeek R1 Distill Llama 70B	won't run	won't run
DeepSeek V3 0324	won't run	won't run
DeepSeek V3.1	won't run	won't run
DeepSeek V3.2	won't run	won't run
DeepSeek V4 Flash 0731	won't run	won't run
DeepSeek V4 Flash Vision Exp	won't run	won't run
Devstral 2 123B Instruct 2512	won't run	won't run
GLM 4.5V	won't run	won't run
GLM 4.6V	won't run	won't run
GLM 4.7	won't run	won't run
GLM 5.1	won't run	won't run
GLM 5.2	won't run	won't run
GLM 5.3	won't run	won't run
Kimi K2	won't run	won't run
Kimi K2 Instruct 0905	won't run	won't run
Kimi K2 Thinking	won't run	won't run
Kimi K2.5	won't run	won't run
MiniMax M2	won't run	won't run
MiniMax M2.1	won't run	won't run
MiniMax M2.5	won't run	won't run
MiniMax M3	won't run	won't run
Mistral Medium 3.5 128B	won't run	won't run

Questions

Which is better for local AI, the Intel Arc Pro B60 or the NVIDIA GeForce RTX 3090?: Choose the NVIDIA GeForce RTX 3090 for speed. Choose the Intel Arc Pro B60 to spend less.
Which can run bigger models, the Intel Arc Pro B60 or the NVIDIA GeForce RTX 3090?: At 64k context both run the same 40 models we track fully in memory.
Which is faster for local AI, the Intel Arc Pro B60 or the NVIDIA GeForce RTX 3090?: On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 2.5× as fast as the Intel Arc Pro B60 (estimated, median across models).
What is the price difference between the Intel Arc Pro B60 and the NVIDIA GeForce RTX 3090?: In Canada the Intel Arc Pro B60 costs $985 at Amazon and the NVIDIA GeForce RTX 3090 $2,800 at Amazon: the Intel Arc Pro B60 is $1,815 cheaper.

Intel Arc Pro B60 vs NVIDIA GeForce RTX 3090

In short

Specifications

What each runs

Questions

More comparisons