readygpu

NVIDIA GeForce RTX 3090 Ti vs NVIDIA GeForce RTX 4070 Super

For running models locally, priced in Germany. At 64k context the first runs 45 models fully in memory, the second 20.

In short

The NVIDIA GeForce RTX 3090 Ti has more memory for models: 24 GB against 12 GB. At 64k context the NVIDIA GeForce RTX 3090 Ti runs 45 of the models we track fully in memory and the NVIDIA GeForce RTX 4070 Super 20; only the NVIDIA GeForce RTX 3090 Ti holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA GeForce RTX 3090 Ti is about 2.0× as fast as the NVIDIA GeForce RTX 4070 Super (estimated, median across models). In Germany the NVIDIA GeForce RTX 3090 Ti costs 1.799 € at eBay (used) and the NVIDIA GeForce RTX 4070 Super 1.800 € at eBay: the NVIDIA GeForce RTX 3090 Ti is 1 € cheaper.

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

NVIDIA GeForce RTX 3090 Ti

1.799 €

eBay, used (asking)

Add to build
NVIDIA GeForce RTX 4070 Super

1.800 €

eBay

Add to build

Specifications

SpecNVIDIA GeForce RTX 3090 TiNVIDIA GeForce RTX 4070 Super
GPUGeForce RTX 3090 TiGeForce RTX 4070 Super
Memory24 GB GDDR6X12 GB GDDR6X
Memory bus384-bit192-bit
Memory bandwidth1,008 GB/s504 GB/s
FP16 compute80 TFLOPS71 TFLOPS
SoftwareCUDACUDA
Board power450 W220 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 3090 TiNVIDIA GeForce RTX 4070 Super
Qwen 3.8 27B29tok/sQ4_K_Moffload
Qwen 3.6 27B26tok/sQ5_K_Moffload
Qwen 3.6 35B-A3B115tok/sQ4_K_Moffload
Qwen 3.5 27B30tok/sQ4_K_Moffload
Gemma 4 26B-A4B78tok/sUD-Q5_K_Moffload
Qwen 3.5 35B-A3B116tok/sQ4_K_Moffload
GLM 4.7 Flash57tok/sQ4_K_Moffload
Nemotron 3.5 Lightning 30B-A3B141tok/sMXFP4offload
gpt-oss 20B93tok/sMXFP4offload
Gemma 4 12B44tok/sQ8_032tok/sQ5_K_M
Qwen 3.5 9B53tok/sQ8_032tok/sQ6_K
Gemma 3 27B28tok/sQ4_K_Moffload
Ministral 3 14B29tok/sQ6_Koffload
Qwen3 14B27tok/sQ6_Koffload
DeepSeek R1 0528 Qwen3 8B34tok/sQ8_0offload
Qwen3 8B34tok/sQ8_0offload
Phi-4 14B26tok/sQ5_K_Moffload
Qwen 3.5 4B94tok/sQ8_047tok/sQ8_0
Ministral 3 8B34tok/sQ8_0offload
Granite 4.2 8B31tok/sQ8_0offload
Gemma 4 E4B62tok/sQ8_031tok/sQ8_0
LFM2.5 8B-A1B88tok/sQ8_046tok/sUD-Q6_K
Llama 3.1 8B41tok/sQ6_Koffload
Qwen2.5 0.5B468tok/sQ8_0234tok/sQ8_0
Qwen3.5 0.8B382tok/sQ8_0191tok/sQ8_0
Qwen3.5 2B277tok/sQ5_K_M138tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B166tok/sQ8_083tok/sQ8_0
MiniCPM5 2B114tok/sQ8_057tok/sQ8_0
Spark X2.5 4B113tok/smlx-4bit57tok/smlx-4bit
Mellum2.1 12B A2.5B Thinking103tok/sQ8_062tok/sQ5_K_M
FrogNano 4B 260992tok/sQ8_046tok/sQ8_0
LFM2.5 2.6B91tok/sQ8_045tok/sQ8_0
LFM2.5 VL 3B91tok/sQ8_045tok/sQ8_0
Kimi VL A3B Thinking 250673tok/sQ8_0offload
Kimi VL A3B71tok/smlx-8bitoffload
granite vision 4.1 4b70tok/sQ8_035tok/sQ8_0
granite 4.2 3b68tok/sQ8_034tok/sQ8_0
GLM 4 9B 041457tok/smlx-5bit29tok/smlx-5bit
LFM2.5 24B A2B54tok/sQ5_K_Moffload
DeepSeek R1 Distill Qwen 7B53tok/sQ8_031tok/sQ6_K
GLM 4.6V Flash49tok/sQ8_035tok/sQ4_K_M
GLM 4.1V 9B Thinking43tok/smlx-8bitoffload
DeepSeek R1 Distill Llama 8B37tok/sQ8_0offload
Ministral 3 8B Reasoning 251234tok/sQ8_0offload
DeepSeek R1 Distill Qwen 14B27tok/sQ5_K_Moffload
Qwen 3.5 122B-A10Boffloadwon't run
Gemma 4 31Boffloadoffload
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadwon't run
Devstral Small 2 24Boffloadoffload
Qwen3 Next 80B-A3Boffloadwon't run
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
Qwen3.8 27B OBLITERATEDoffloadoffload
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 R1 Zerowon'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 3090 Ti or the NVIDIA GeForce RTX 4070 Super?
Choose the NVIDIA GeForce RTX 3090 Ti to run bigger models. Choose the NVIDIA GeForce RTX 3090 Ti for speed. Choose the NVIDIA GeForce RTX 3090 Ti to spend less.
Which can run bigger models, the NVIDIA GeForce RTX 3090 Ti or the NVIDIA GeForce RTX 4070 Super?
At 64k context the NVIDIA GeForce RTX 3090 Ti runs 45 of the models we track fully in memory and the NVIDIA GeForce RTX 4070 Super 20; only the NVIDIA GeForce RTX 3090 Ti holds Qwen 3.8 27B.
Which is faster for local AI, the NVIDIA GeForce RTX 3090 Ti or the NVIDIA GeForce RTX 4070 Super?
On the models both run fully in memory, the NVIDIA GeForce RTX 3090 Ti is about 2.0× as fast as the NVIDIA GeForce RTX 4070 Super (estimated, median across models).
What is the price difference between the NVIDIA GeForce RTX 3090 Ti and the NVIDIA GeForce RTX 4070 Super?
In Germany the NVIDIA GeForce RTX 3090 Ti costs 1.799 € at eBay (used) and the NVIDIA GeForce RTX 4070 Super 1.800 € at eBay: the NVIDIA GeForce RTX 3090 Ti is 1 € cheaper.

More comparisons