readygpu

NVIDIA GeForce RTX 3090 vs NVIDIA GeForce RTX 4070 Ti Super

For running models locally, priced in India. At 32k context the first runs 26 models fully in memory, the second 13.

In short

The NVIDIA GeForce RTX 3090 has more memory for models: 24 GB against 16 GB. At 32k context the NVIDIA GeForce RTX 3090 runs 26 of the models we track fully in memory and the NVIDIA GeForce RTX 4070 Ti Super 13; only the NVIDIA GeForce RTX 3090 holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 1.4× as fast as the NVIDIA GeForce RTX 4070 Ti Super (estimated, median across models).

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

NVIDIA GeForce RTX 3090

Not priced here

Add to build
NVIDIA GeForce RTX 4070 Ti Super

Not priced here

Add to build

Specifications

SpecNVIDIA GeForce RTX 3090NVIDIA GeForce RTX 4070 Ti Super
GPUGeForce RTX 3090GeForce RTX 4070 Ti Super
Memory24 GB GDDR6X16 GB GDDR6X
Memory bus384-bit256-bit
Memory bandwidth936 GB/s672 GB/s
FP16 compute71 TFLOPS88 TFLOPS
SoftwareCUDACUDA
Board power350 W285 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 3090NVIDIA GeForce RTX 4070 Ti Super
Qwen 3.8 27B26tok/sUD-Q5_K_Moffload
Qwen 3.6 27B27tok/sQ5_K_Moffload
Gemma 4 31B27tok/sQ4_K_Moffload
Qwen 3.6 35B-A3B135tok/sUD-Q4_K_Moffload
Qwen 3.5 27B27tok/sQ5_K_Moffload
Gemma 4 26B-A4B78tok/sUD-Q6_Koffload
Qwen 3.5 35B-A3B135tok/sQ4_K_Moffload
GLM 4.7 Flash65tok/sQ5_K_Moffload
Devstral Small 2 24B26tok/sQ5_K_Moffload
gpt-oss 20B108tok/sMXFP477tok/sMXFP4
Qwen3 Coder 30B-A3B69tok/sQ4_K_Moffload
Gemma 4 12B43tok/sQ8_031tok/sQ8_0
Qwen 3.5 9B55tok/sQ8_039tok/sQ8_0
Mistral Small 3.2 24B26tok/sQ5_K_Moffload
Gemma 3 27B26tok/sQ5_K_Moffload
Ministral 3 14B29tok/sQ8_028tok/sQ5_K_M
Qwen3 14B28tok/sQ8_029tok/sQ4_K_M
DeepSeek R1 0528 Qwen3 8B43tok/sQ8_031tok/sQ8_0
Qwen3 8B43tok/sQ8_031tok/sQ8_0
Phi-4 14B26tok/sQ8_0offload
Qwen 3.5 4B104tok/sQ8_074tok/sQ8_0
Ministral 3 8B43tok/sQ8_031tok/sQ8_0
Granite 4.2 8B39tok/sQ8_028tok/sQ8_0
Gemma 4 E4B63tok/sQ8_045tok/sQ8_0
LFM2.5 8B-A1B130tok/sQ8_094tok/sQ8_0
Llama 3.1 8B53tok/sQ6_K38tok/sQ6_K
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
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

Questions

Which is better for local AI, the NVIDIA GeForce RTX 3090 or the NVIDIA GeForce RTX 4070 Ti Super?
Choose the NVIDIA GeForce RTX 3090 to run bigger models. Choose the NVIDIA GeForce RTX 3090 for speed.
Which can run bigger models, the NVIDIA GeForce RTX 3090 or the NVIDIA GeForce RTX 4070 Ti Super?
At 32k context the NVIDIA GeForce RTX 3090 runs 26 of the models we track fully in memory and the NVIDIA GeForce RTX 4070 Ti Super 13; only the NVIDIA GeForce RTX 3090 holds Qwen 3.8 27B.
Which is faster for local AI, the NVIDIA GeForce RTX 3090 or the NVIDIA GeForce RTX 4070 Ti Super?
On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 1.4× as fast as the NVIDIA GeForce RTX 4070 Ti Super (estimated, median across models).

More comparisons