NVIDIA RTX 3090 Ti vs NVIDIA RTX 4090 for local LLMs
6 reports on the NVIDIA RTX 3090 Ti and 57 on the NVIDIA RTX 4090
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 about the same on both (1,008 GB/s on the NVIDIA RTX 3090 Ti, 1,008 on the NVIDIA RTX 4090). The estimates below compare them on the models people run most.
NVIDIA RTX 3090 Ti
- Memory
- 24GB
- Memory bandwidth
- 1,008 GB/s
- FP16 compute
- 160 TFLOPS
- Reports
- 6
NVIDIA RTX 4090
- Memory
- 24GB
- Memory bandwidth
- 1,008 GB/s
- FP16 compute
- 330 TFLOPS
- Reports
- 57
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.
| Model | NVIDIA RTX 3090 Ti | NVIDIA RTX 4090 |
|---|---|---|
| Qwen3.8 27B | 36.5 t/sQ4_K_M | 36.5 t/sQ4_K_M |
| Qwen3.8 125B · 6B active | 19.6 t/sQ2_K, experts in RAM | 19.6 t/sQ2_K, experts in RAM |
| Qwen3.6 35B · 3B active | 129.3 t/sQ4_K_S | 129.3 t/sQ4_K_S |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 27B | 36.5 t/sQ4_K_M | 36.5 t/sQ4_K_M |
| Gemma 4 26B · 4B active | 82.3 t/sQ5_K_S | 82.3 t/sQ5_K_S |
Change the context, RAM or model in the calculator: NVIDIA RTX 3090 Ti · NVIDIA RTX 4090
Every report on each card, with its source: NVIDIA RTX 3090 Ti · NVIDIA RTX 4090