readygpu

NVIDIA GeForce RTX 4090 vs AMD Radeon RX 7900 XT

For running models locally, priced in the US. At 8k context the first runs 57 models fully in memory, the second 49.

In short

The NVIDIA GeForce RTX 4090 has more memory for models: 24 GB against 20 GB. At 8k context the NVIDIA GeForce RTX 4090 runs 57 of the models we track fully in memory and the AMD Radeon RX 7900 XT 49; only the NVIDIA GeForce RTX 4090 holds Gemma 4 31B. On the models both run fully in memory, the NVIDIA GeForce RTX 4090 is about 1.4× as fast as the AMD Radeon RX 7900 XT (estimated, median across models). In the US the NVIDIA GeForce RTX 4090 costs $2,073 at eBay and the AMD Radeon RX 7900 XT $800 at Amazon: the AMD Radeon RX 7900 XT is $1,274 cheaper.

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

NVIDIA GeForce RTX 4090

$2,073

eBay

Add to build
AMD Radeon RX 7900 XT

$800

Amazon

Add to build

Specifications

SpecNVIDIA GeForce RTX 4090AMD Radeon RX 7900 XT
GPUGeForce RTX 4090Radeon RX 7900 XT
Memory24 GB GDDR6X20 GB GDDR6
Memory bus384-bit320-bit
Memory bandwidth1,008 GB/s800 GB/s
FP16 compute165 TFLOPS103 TFLOPS
SoftwareCUDAROCm
Board power450 W315 W
PCIe generationGen 4Gen 4
PCIe lanesx16x16

What each runs

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

ModelNVIDIA GeForce RTX 4090AMD Radeon RX 7900 XT
Qwen 3.8 27B28tok/sUD-Q6_K24tok/sQ4_K_M
Qwen 3.6 27B27tok/sQ6_K25tok/sQ4_K_M
Gemma 4 31B27tok/sQ5_K_Moffload
Qwen 3.6 35B-A3B179tok/sQ4_K_Moffload
Qwen 3.5 27B27tok/sQ6_K25tok/sQ4_K_M
Gemma 4 26B-A4B95tok/sUD-Q6_K78tok/sQ5_K_M
Qwen 3.5 35B-A3B181tok/sQ4_K_Moffload
GLM 4.7 Flash94tok/sQ5_K_M76tok/sQ4_K_M
Devstral Small 2 24B30tok/sQ6_K24tok/sQ5_K_M
Nemotron 3.5 Lightning 30B-A3B163tok/sMXFP4offload
gpt-oss 20B142tok/sMXFP4100tok/sMXFP4
Qwen3 Coder 30B-A3B128tok/sQ5_K_M100tok/sQ4_K_M
Qwen3 32B29tok/sQ4_K_Moffload
Gemma 4 12B48tok/sQ8_034tok/sQ8_0
Qwen 3.5 9B64tok/sQ8_045tok/sQ8_0
Mistral Small 3.2 24B30tok/sQ6_K24tok/sQ5_K_M
DeepSeek R1 Distill Qwen 32B28tok/sQ4_K_Moffload
Gemma 3 27B27tok/sQ6_K25tok/sQ4_K_M
Ministral 3 14B40tok/sQ8_028tok/sQ8_0
Qwen3 14B37tok/sQ8_026tok/sQ8_0
DeepSeek R1 0528 Qwen3 8B63tok/sQ8_044tok/sQ8_0
Qwen3 8B63tok/sQ8_044tok/sQ8_0
Phi-4 14B36tok/sQ8_026tok/sQ8_0
Qwen 3.5 4B130tok/sQ8_092tok/sQ8_0
Ministral 3 8B62tok/sQ8_043tok/sQ8_0
Granite 4.2 8B59tok/sQ8_041tok/sQ8_0
Gemma 4 E4B74tok/sQ8_052tok/sQ8_0
LFM2.5 8B-A1B256tok/sQ8_0180tok/sQ8_0
Llama 3.1 8B82tok/sQ6_K57tok/sQ6_K
Qwen2.5 0.5B989tok/sQ8_0697tok/sQ8_0
Qwen3.5 0.8B671tok/sQ8_0472tok/sQ8_0
Qwen3.5 2B402tok/sQ5_K_M283tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B294tok/sQ8_0207tok/sQ8_0
MiniCPM5 2B206tok/sQ8_0145tok/sQ8_0
LFM2.5 2.6B185tok/sQ8_0130tok/sQ8_0
LFM2.5 VL 3B185tok/sQ8_0130tok/sQ8_0
Spark X2.5 4B184tok/smlx-4bit130tok/smlx-4bit
LFM2.5 24B A2B154tok/sQ6_K119tok/sQ5_K_M
granite vision 4.1 4b146tok/sQ8_0103tok/sQ8_0
granite 4.2 3b137tok/sQ8_096tok/sQ8_0
Mellum2.1 12B A2.5B Thinking133tok/sQ8_094tok/sQ8_0
FrogNano 4B 2609127tok/sQ8_089tok/sQ8_0
Kimi VL A3B Thinking 2506112tok/sQ8_079tok/sQ8_0
Kimi VL A3B107tok/smlx-8bit75tok/smlx-8bit
DeepSeek R1 Distill Qwen 7B73tok/sQ8_051tok/sQ8_0
GLM 4 9B 041473tok/smlx-5bit51tok/smlx-5bit
DeepSeek R1 Distill Llama 8B65tok/sQ8_046tok/sQ8_0
Ministral 3 8B Reasoning 251262tok/sQ8_043tok/sQ8_0
GLM 4.6V Flash61tok/sQ8_043tok/sQ8_0
GLM 4.1V 9B Thinking52tok/smlx-8bit36tok/smlx-8bit
DeepSeek R1 Distill Qwen 14B36tok/sQ8_025tok/sQ8_0
Devstral Small 250735tok/sQ5_K_M24tok/sQ5_K_M
Magistral Small 250735tok/sQ5_K_M24tok/sQ5_K_M
Magistral Small 250935tok/sQ5_K_M24tok/sQ5_K_M
Magistral Small 250630tok/sQ6_K24tok/sQ5_K_M
Qwen3.8 27B OBLITERATED29tok/smlx-4bitoffload
granite 4.2 30b27tok/sQ5_K_Moffload
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadoffload
GLM 4.5 Airoffloadoffload
Qwen3 Next 80B-A3Boffloadwon't run
Kimi Linear 48B-A3Boffloadoffload
Llama 3.3 70Boffloadoffload
DeepSeek R1 Distill Llama 70Boffloadoffload
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
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Zerowon't runwon't run
DeepSeek V3won't runwon't run
DeepSeek V3 0324won't runwon't run
DeepSeek V3.1won't runwon't run
DeepSeek V3.1 Terminuswon'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 4.5won't runwon't run
GLM 4.5Vwon't runwon't run
GLM 4.6won't runwon't run
GLM 4.6Vwon't runwon't run
GLM 4.7won't runwon't run
GLM 5won'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 Dev 72Bwon'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 AMD Radeon RX 7900 XT?
Choose the NVIDIA GeForce RTX 4090 to run bigger models. Choose the NVIDIA GeForce RTX 4090 for speed. Choose the AMD Radeon RX 7900 XT to spend less.
Which can run bigger models, the NVIDIA GeForce RTX 4090 or the AMD Radeon RX 7900 XT?
At 8k context the NVIDIA GeForce RTX 4090 runs 57 of the models we track fully in memory and the AMD Radeon RX 7900 XT 49; only the NVIDIA GeForce RTX 4090 holds Gemma 4 31B.
Which is faster for local AI, the NVIDIA GeForce RTX 4090 or the AMD Radeon RX 7900 XT?
On the models both run fully in memory, the NVIDIA GeForce RTX 4090 is about 1.4× as fast as the AMD Radeon RX 7900 XT (estimated, median across models).
What is the price difference between the NVIDIA GeForce RTX 4090 and the AMD Radeon RX 7900 XT?
In the US the NVIDIA GeForce RTX 4090 costs $2,073 at eBay and the AMD Radeon RX 7900 XT $800 at Amazon: the AMD Radeon RX 7900 XT is $1,274 cheaper.

More comparisons