llamaperf

NVIDIA RTX 2000 Ada vs NVIDIA RTX 3090 for local LLMs

2 reports on the NVIDIA RTX 2000 Ada and 271 on the NVIDIA RTX 3090

Which is faster for local LLMs?

Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 4.2x higher on the NVIDIA RTX 3090 (224 GB/s on the NVIDIA RTX 2000 Ada, 936 on the NVIDIA RTX 3090). The estimates below compare them on the models people run most.

NVIDIA RTX 2000 Ada
Memory
16GB
Memory bandwidth
224 GB/s
FP16 compute
48 TFLOPS
Reports
2
NVIDIA RTX 3090
Memory
24GB
Memory bandwidth
936 GB/s
FP16 compute
142 TFLOPS
Reports
271

Estimated by the calculator

The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape.

ModelNVIDIA RTX 2000 AdaNVIDIA RTX 3090
Qwen3.8 27B14.6 t/sQ2_K33.9 t/sQ4_K_M
Qwen3.8 125B · 6B activedoesn't fit at these settings19.4 t/sQ2_K, experts in RAM
Qwen3.6 35B · 3B active48.3 t/sQ2_K120.1 t/sQ4_K_S
DeepSeek V4 Flash 284B · 13B activedoesn't fit at these settingsdoesn't fit at these settings
DeepSeek V4.1 Flash 552B · 16B activedoesn't fit at these settingsdoesn't fit at these settings
Qwen3.6 27B14.6 t/sQ2_K33.9 t/sQ4_K_M

Change the context, RAM or model in the calculator: NVIDIA RTX 2000 Ada · NVIDIA RTX 3090

Every report on each card, with its source: NVIDIA RTX 2000 Ada · NVIDIA RTX 3090

More comparisons