readygpu

Intel Arc B580 vs NVIDIA GeForce RTX 4090

For running models locally, priced in Canada. At 64k context the first runs 16 models fully in memory, the second 40.

In short

The NVIDIA GeForce RTX 4090 has more memory for models: 24 GB against 12 GB. At 64k context the NVIDIA GeForce RTX 4090 runs 40 of the models we track fully in memory and the Intel Arc B580 16; only the NVIDIA GeForce RTX 4090 holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA GeForce RTX 4090 is about 2.7× as fast as the Intel Arc B580 (estimated, median across models). In Canada the Intel Arc B580 costs $550 at eBay and the NVIDIA GeForce RTX 4090 $3,300 at eBay: the Intel Arc B580 is $2,750 cheaper.

Updated . Speeds are estimates; how they are worked out.

Intel Arc B580

$550

eBay

Add to build
NVIDIA GeForce RTX 4090

$3,300

eBay

Add to build

Specifications

SpecIntel Arc B580NVIDIA GeForce RTX 4090
GPUArc B580GeForce RTX 4090
Memory12 GB GDDR624 GB GDDR6X
Memory bus192-bit384-bit
Memory bandwidth456 GB/s1,008 GB/s
FP16 compute116.5 TFLOPS165 TFLOPS
SoftwareSYCL / oneAPICUDA
Board power190 W450 W
PCIe generationGen 4Gen 4
PCIe lanesx8x16

What each runs

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

ModelIntel Arc B580NVIDIA GeForce RTX 4090
Qwen 3.8 27Boffload29tok/sQ4_K_M
Qwen 3.6 27Boffload26tok/sQ5_K_M
Qwen 3.6 35B-A3Boffload115tok/sQ4_K_M
Qwen 3.5 27Boffload30tok/sQ4_K_M
Gemma 4 26B-A4Boffload78tok/sUD-Q5_K_M
Qwen 3.5 35B-A3Boffload116tok/sQ4_K_M
GLM 4.7 Flashoffload57tok/sQ4_K_M
gpt-oss 20Boffload93tok/sMXFP4
Gemma 4 12B23tok/sQ5_K_M44tok/sQ8_0
Qwen 3.5 9B24tok/sQ6_K53tok/sQ8_0
Gemma 3 27Boffload28tok/sQ4_K_M
Ministral 3 14Boffload29tok/sQ6_K
Qwen3 14Boffload27tok/sQ6_K
DeepSeek R1 0528 Qwen3 8Boffload34tok/sQ8_0
Qwen3 8Boffload34tok/sQ8_0
Phi-4 14Boffload26tok/sQ5_K_M
Qwen 3.5 4B34tok/sQ8_094tok/sQ8_0
Ministral 3 8Boffload34tok/sQ8_0
Granite 4.2 8Boffload31tok/sQ8_0
Gemma 4 E4B23tok/sQ8_062tok/sQ8_0
LFM2.5 8B-A1B31tok/sUD-Q6_K88tok/sQ8_0
Llama 3.1 8Boffload41tok/sQ6_K
Qwen2.5 0.5B171tok/sQ8_0468tok/sQ8_0
Qwen3.5 0.8B139tok/sQ8_0382tok/sQ8_0
Qwen3.5 2B101tok/sQ5_K_M277tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B60tok/sQ8_0166tok/sQ8_0
MiniCPM5 2B42tok/sQ8_0114tok/sQ8_0
LFM2.5 2.6B33tok/sQ8_091tok/sQ8_0
LFM2.5 VL 3B33tok/sQ8_091tok/sQ8_0
GLM 4.6V Flash26tok/sQ4_K_M49tok/sQ8_0
granite vision 4.1 4b25tok/sQ8_070tok/sQ8_0
granite 4.2 3b25tok/sQ8_068tok/sQ8_0
DeepSeek R1 Distill Qwen 7B23tok/sQ6_K53tok/sQ8_0
Mellum2.1 12B A2.5B Thinkingoffload103tok/sQ8_0
Kimi VL A3Boffload71tok/smlx-8bit
DeepSeek R1 Distill Llama 8Boffload37tok/sQ8_0
Ministral 3 8B Reasoning 2512offload34tok/sQ8_0
GLM 4.1V 9B Thinkingoffload43tok/smlx-8bit
LFM2.5 24B A2Boffload54tok/sQ5_K_M
DeepSeek R1 Distill Qwen 14Boffload27tok/sQ5_K_M
Qwen 3.5 122B-A10Bwon't runoffload
Gemma 4 31Boffloadoffload
Qwen3 Coder Nextwon't runoffload
Mistral Small 4 119Bwon't runoffload
Devstral Small 2 24Boffloadoffload
Qwen3 Next 80B-A3Bwon't runoffload
Nemotron 3.5 Lightning 30B-A3Boffloadoffload
Kimi Linear 48B-A3Boffloadoffload
Qwen3 Coder 30B-A3Boffloadoffload
Qwen3 32Boffloadoffload
Mistral Small 3.2 24Boffloadoffload
DeepSeek R1 Distill Qwen 32Boffloadoffload
Devstral Small 2507offloadoffload
Magistral Small 2506offloadoffload
Magistral Small 2507offloadoffload
Magistral Small 2509offloadoffload
granite 4.2 30boffloadoffload
DeepSeek V4 Flashwon't runwon't run
GLM 5.3 Flashwon't runwon't run
Qwen 3.8 Flash Nextwon't runwon't run
MiniMax M2.7won't runwon't run
Nemotron 3 Super 120B-A12Bwon't runwon't run
gpt-oss 120Bwon't runwon't run
GLM 4.5 Airwon't runwon't run
Llama 3.3 70Bwon't runwon't run
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Distill Llama 70Bwon't runwon't run
DeepSeek V3 0324won't runwon't run
DeepSeek V3.1won't runwon't run
DeepSeek V3.2won't runwon't run
DeepSeek V4 Flash 0731won't runwon't run
DeepSeek V4 Flash Vision Expwon't runwon't run
Devstral 2 123B Instruct 2512won't runwon't run
GLM 4.5Vwon't runwon't run
GLM 4.6Vwon't runwon't run
GLM 4.7won't runwon't run
GLM 5.1won't runwon't run
GLM 5.2won't runwon't run
GLM 5.3won't runwon't run
Kimi K2won't runwon't run
Kimi K2 Instruct 0905won't runwon't run
Kimi K2 Thinkingwon't runwon't run
Kimi K2.5won't runwon't run
MiniMax M2won't runwon't run
MiniMax M2.1won't runwon't run
MiniMax M2.5won't runwon't run
MiniMax M3won't runwon't run
Mistral Medium 3.5 128Bwon't runwon't run

Questions

Which is better for local AI, the Intel Arc B580 or the NVIDIA GeForce RTX 4090?
Choose the NVIDIA GeForce RTX 4090 to run bigger models. Choose the NVIDIA GeForce RTX 4090 for speed. Choose the Intel Arc B580 to spend less.
Which can run bigger models, the Intel Arc B580 or the NVIDIA GeForce RTX 4090?
At 64k context the NVIDIA GeForce RTX 4090 runs 40 of the models we track fully in memory and the Intel Arc B580 16; only the NVIDIA GeForce RTX 4090 holds Qwen 3.8 27B.
Which is faster for local AI, the Intel Arc B580 or the NVIDIA GeForce RTX 4090?
On the models both run fully in memory, the NVIDIA GeForce RTX 4090 is about 2.7× as fast as the Intel Arc B580 (estimated, median across models).
What is the price difference between the Intel Arc B580 and the NVIDIA GeForce RTX 4090?
In Canada the Intel Arc B580 costs $550 at eBay and the NVIDIA GeForce RTX 4090 $3,300 at eBay: the Intel Arc B580 is $2,750 cheaper.

More comparisons