readygpu

NVIDIA GeForce RTX 4090 vs NVIDIA GeForce RTX 5050

For running models locally, priced in Canada. At 32k context the first runs 58 models fully in memory, the second 18.

In short

The NVIDIA GeForce RTX 4090 has more memory for models: 24 GB against 8 GB. At 32k context the NVIDIA GeForce RTX 4090 runs 58 of the models we track fully in memory and the NVIDIA GeForce RTX 5050 18; 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 3.1× as fast as the NVIDIA GeForce RTX 5050 (estimated, median across models). In Canada the NVIDIA GeForce RTX 4090 costs $8,402 at Amazon and the NVIDIA GeForce RTX 5050 $570 at Amazon: the NVIDIA GeForce RTX 5050 is $7,832 cheaper.

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

NVIDIA GeForce RTX 4090

$8,402

Amazon

Add to build
NVIDIA GeForce RTX 5050

$570

Amazon

Add to build

Specifications

SpecNVIDIA GeForce RTX 4090NVIDIA GeForce RTX 5050
GPUGeForce RTX 4090GeForce RTX 5050
Memory24 GB GDDR6X8 GB GDDR6
Memory bus384-bit128-bit
Memory bandwidth1,008 GB/s320 GB/s
FP16 compute165 TFLOPS13.169 TFLOPS
SoftwareCUDACUDA
Board power450 W130 W
PCIe generationGen 4Gen 5
PCIe lanesx16x8

What each runs

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

ModelNVIDIA GeForce RTX 4090NVIDIA GeForce RTX 5050
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
Nemotron 3.5 Lightning 30B-A3B153tok/sMXFP4offload
gpt-oss 20B116tok/sMXFP4offload
Qwen3 Coder 30B-A3B74tok/sQ4_K_Moffload
Gemma 4 12B46tok/sQ8_0offload
Qwen 3.5 9B59tok/sQ8_029tok/sQ4_K_M
Mistral Small 3.2 24B28tok/sQ5_K_Moffload
Gemma 3 27B28tok/sQ5_K_Moffload
Ministral 3 14B32tok/sQ8_0offload
Qwen3 14B30tok/sQ8_0offload
DeepSeek R1 0528 Qwen3 8B46tok/sQ8_0offload
Qwen3 8B46tok/sQ8_0offload
Phi-4 14B28tok/sQ8_0offload
Qwen 3.5 4B112tok/sQ8_035tok/sQ8_0
Ministral 3 8B46tok/sQ8_0offload
Granite 4.2 8B43tok/sQ8_0offload
Gemma 4 E4B68tok/sQ8_031tok/sQ5_K_M
LFM2.5 8B-A1B140tok/sQ8_053tok/sMXFP4
Llama 3.1 8B57tok/sQ6_Koffload
Qwen2.5 0.5B669tok/sQ8_0213tok/sQ8_0
Qwen3.5 0.8B507tok/sQ8_0161tok/sQ8_0
Qwen3.5 2B337tok/sQ5_K_M107tok/sQ5_K_M
LFM3 1.7B277tok/sQ8_088tok/sQ8_0
DeepSeek R1 Distill Qwen 1.5B221tok/sQ8_070tok/sQ8_0
MiniCPM5 2B153tok/sQ8_049tok/sQ8_0
Spark X2.5 4B145tok/smlx-4bit46tok/smlx-4bit
LFM2.5 2.6B128tok/sQ8_041tok/sQ8_0
LFM2.5 VL 3B128tok/sQ8_041tok/sQ8_0
Mellum2.1 12B A2.5B Thinking118tok/sQ8_0offload
FrogNano 4B 2609109tok/sQ8_035tok/sQ8_0
granite vision 4.1 4b99tok/sQ8_032tok/sQ8_0
granite 4.2 3b95tok/sQ8_030tok/sQ8_0
Kimi VL A3B Thinking 250691tok/sQ8_0offload
Kimi VL A3B88tok/smlx-8bitoffload
Ministral 3 3B Instruct 251288tok/sQ8_031tok/sQ6_K
LFM2.5 24B A2B84tok/sQ6_Koffload
GLM 4 9B 041465tok/smlx-5bitoffload
DeepSeek R1 Distill Qwen 7B63tok/sQ8_030tok/sQ4_K_M
GLM 4.6V Flash55tok/sQ8_0offload
DeepSeek R1 Distill Llama 8B49tok/sQ8_0offload
GLM 4.1V 9B Thinking48tok/smlx-8bitoffload
Ministral 3 8B Reasoning 251246tok/sQ8_0offload
Ministral 3 14B Reasoning 251232tok/sQ8_0offload
DeepSeek R1 Distill Qwen 14B28tok/sQ8_0offload
Devstral Small 250728tok/sQ5_K_Moffload
Magistral Small 250628tok/sQ5_K_Moffload
Magistral Small 250728tok/sQ5_K_Moffload
Magistral Small 250928tok/sQ5_K_Moffload
Qwen3.8 27B OBLITERATED27tok/smlx-4bitoffload
Muse Glimmer 30B26tok/sQ6_Koffload
Qwen 3.5 122B-A10Boffloadwon't run
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadwon't run
GLM 4.5 Airoffloadwon't run
Qwen3 Next 80B-A3Boffloadwon't run
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 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 NVIDIA GeForce RTX 5050?
Choose the NVIDIA GeForce RTX 4090 to run bigger models. Choose the NVIDIA GeForce RTX 4090 for speed. Choose the NVIDIA GeForce RTX 5050 to spend less.
Which can run bigger models, the NVIDIA GeForce RTX 4090 or the NVIDIA GeForce RTX 5050?
At 32k context the NVIDIA GeForce RTX 4090 runs 58 of the models we track fully in memory and the NVIDIA GeForce RTX 5050 18; 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 5050?
On the models both run fully in memory, the NVIDIA GeForce RTX 4090 is about 3.1× as fast as the NVIDIA GeForce RTX 5050 (estimated, median across models).
What is the price difference between the NVIDIA GeForce RTX 4090 and the NVIDIA GeForce RTX 5050?
In Canada the NVIDIA GeForce RTX 4090 costs $8,402 at Amazon and the NVIDIA GeForce RTX 5050 $570 at Amazon: the NVIDIA GeForce RTX 5050 is $7,832 cheaper.

More comparisons

More hardware for local AI