readygpu

NVIDIA GeForce RTX 4090 vs NVIDIA GeForce RTX 5080

For running models locally, priced in Canada. At 32k context the first runs 48 models fully in memory, the second 30.

In short

The NVIDIA GeForce RTX 4090 has more memory for models: 24 GB against 16 GB. At 32k context the NVIDIA GeForce RTX 4090 runs 48 of the models we track fully in memory and the NVIDIA GeForce RTX 5080 30; only the NVIDIA GeForce RTX 4090 holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA GeForce RTX 4090 and the NVIDIA GeForce RTX 5080 are about as fast as each other (estimated). In Canada the NVIDIA GeForce RTX 4090 costs $3,300 at eBay and the NVIDIA GeForce RTX 5080 $2,370 at Amazon: the NVIDIA GeForce RTX 5080 is $930 cheaper.

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

NVIDIA GeForce RTX 4090

$3,300

eBay

Add to build
NVIDIA GeForce RTX 5080

$2,370

Amazon

Add to build

Specifications

SpecNVIDIA GeForce RTX 4090NVIDIA GeForce RTX 5080
GPUGeForce RTX 4090GeForce RTX 5080
Memory24 GB GDDR6X16 GB GDDR7
Memory bus384-bit256-bit
Memory bandwidth1,008 GB/s960 GB/s
FP16 compute165 TFLOPS113 TFLOPS
SoftwareCUDACUDA
Board power450 W360 W
PCIe generationGen 4Gen 5
PCIe lanesx16x16

What each runs

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

ModelNVIDIA GeForce RTX 4090NVIDIA GeForce RTX 5080
Qwen 3.8 27B27tok/sQ5_K_Moffload
Qwen 3.6 27B29tok/sQ5_K_Moffload
Gemma 4 31B29tok/sQ4_K_Moffload
Qwen 3.6 35B-A3B145tok/sQ4_K_Moffload
Qwen 3.5 27B29tok/sQ5_K_Moffload
Gemma 4 26B-A4B84tok/sUD-Q6_Koffload
Qwen 3.5 35B-A3B146tok/sQ4_K_Moffload
GLM 4.7 Flash70tok/sQ5_K_Moffload
Devstral Small 2 24B28tok/sQ5_K_Moffload
gpt-oss 20B116tok/sMXFP4110tok/sMXFP4
Qwen3 Coder 30B-A3B74tok/sQ4_K_Moffload
Gemma 4 12B46tok/sQ8_044tok/sQ8_0
Qwen 3.5 9B59tok/sQ8_056tok/sQ8_0
Mistral Small 3.2 24B28tok/sQ5_K_Moffload
Gemma 3 27B28tok/sQ5_K_Moffload
Ministral 3 14B32tok/sQ8_040tok/sQ5_K_M
Qwen3 14B30tok/sQ8_041tok/sQ4_K_M
DeepSeek R1 0528 Qwen3 8B46tok/sQ8_044tok/sQ8_0
Qwen3 8B46tok/sQ8_044tok/sQ8_0
Phi-4 14B28tok/sQ8_0offload
Qwen 3.5 4B112tok/sQ8_0106tok/sQ8_0
Ministral 3 8B46tok/sQ8_044tok/sQ8_0
Granite 4.2 8B43tok/sQ8_041tok/sQ8_0
Gemma 4 E4B68tok/sQ8_065tok/sQ8_0
LFM2.5 8B-A1B140tok/sQ8_0134tok/sQ8_0
Llama 3.1 8B57tok/sQ6_K55tok/sQ6_K
Qwen2.5 0.5B669tok/sQ8_0637tok/sQ8_0
Qwen3.5 0.8B507tok/sQ8_0483tok/sQ8_0
Qwen3.5 2B337tok/sQ5_K_M321tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B221tok/sQ8_0210tok/sQ8_0
MiniCPM5 2B153tok/sQ8_0146tok/sQ8_0
LFM2.5 2.6B128tok/sQ8_0122tok/sQ8_0
LFM2.5 VL 3B128tok/sQ8_0122tok/sQ8_0
Mellum2.1 12B A2.5B Thinking118tok/sQ8_0113tok/sQ8_0
granite vision 4.1 4b99tok/sQ8_094tok/sQ8_0
granite 4.2 3b95tok/sQ8_091tok/sQ8_0
Kimi VL A3B88tok/smlx-8bit101tok/smlx-6bit
LFM2.5 24B A2B84tok/sQ6_Koffload
DeepSeek R1 Distill Qwen 7B63tok/sQ8_060tok/sQ8_0
GLM 4.6V Flash55tok/sQ8_053tok/sQ8_0
DeepSeek R1 Distill Llama 8B49tok/sQ8_046tok/sQ8_0
GLM 4.1V 9B Thinking48tok/smlx-8bit45tok/smlx-8bit
Ministral 3 8B Reasoning 251246tok/sQ8_044tok/sQ8_0
DeepSeek R1 Distill Qwen 14B28tok/sQ8_039tok/sQ4_K_M
Devstral Small 250728tok/sQ5_K_Moffload
Magistral Small 250628tok/sQ5_K_Moffload
Magistral Small 250728tok/sQ5_K_Moffload
Magistral Small 250928tok/sQ5_K_Moffload
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadoffload
GLM 4.5 Airoffloadwon't run
Qwen3 Next 80B-A3Boffloadwon't run
Nemotron 3.5 Lightning 30B-A3Boffloadoffload
Kimi Linear 48B-A3Boffloadoffload
Qwen3 32Boffloadoffload
DeepSeek R1 Distill Qwen 32Boffloadoffload
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
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 NVIDIA GeForce RTX 4090 or the NVIDIA GeForce RTX 5080?
Choose the NVIDIA GeForce RTX 4090 to run bigger models. Choose the NVIDIA GeForce RTX 5080 to spend less.
Which can run bigger models, the NVIDIA GeForce RTX 4090 or the NVIDIA GeForce RTX 5080?
At 32k context the NVIDIA GeForce RTX 4090 runs 48 of the models we track fully in memory and the NVIDIA GeForce RTX 5080 30; only the NVIDIA GeForce RTX 4090 holds Qwen 3.8 27B.
Which is faster for local AI, the NVIDIA GeForce RTX 4090 or the NVIDIA GeForce RTX 5080?
On the models both run fully in memory, the NVIDIA GeForce RTX 4090 and the NVIDIA GeForce RTX 5080 are about as fast as each other (estimated).
What is the price difference between the NVIDIA GeForce RTX 4090 and the NVIDIA GeForce RTX 5080?
In Canada the NVIDIA GeForce RTX 4090 costs $3,300 at eBay and the NVIDIA GeForce RTX 5080 $2,370 at Amazon: the NVIDIA GeForce RTX 5080 is $930 cheaper.

More comparisons