RTX 5080 vs RTX 4090 for local LLMs

The RTX 4090 runs more models well (48 vs 28 of 69), thanks to its 24 GB of memory. On models that fit both, the RTX 4090 should be about 5% faster, because generation speed follows memory bandwidth (1008 vs 960 GB/s).

NVIDIA RTX 5080

  • 16 GB VRAM
  • 960 GB/s bandwidth
  • 28/69 models run well
  • $999 at launch

NVIDIA RTX 4090

  • 24 GB VRAM
  • 1008 GB/s bandwidth
  • 48/69 models run well
  • $1,599 at launch

Prices are US launch prices (MSRP) for comparison; street prices change daily. Links go to Amazon, where we earn a commission at no cost to you.

Every model, side by side

Best quant that fits and estimated speed at 8K context.

ModelRTX 5080RTX 4090
Qwen3 0.6BFP16 · ~384 tok/sFP16 · ~403 tok/s
Gemma 3 1BFP16 · ~288 tok/sFP16 · ~302 tok/s
Llama 3.2 1BFP16 · ~232 tok/sFP16 · ~244 tok/s
Qwen3 1.7BFP16 · ~142 tok/sFP16 · ~149 tok/s
Llama 3.2 3BFP16 · ~90 tok/sFP16 · ~94 tok/s
Phi-3.5 Mini 3.8BFP16 · ~75 tok/sFP16 · ~79 tok/s
Phi-4 Mini 3.8BFP16 · ~75 tok/sFP16 · ~79 tok/s
Qwen3 4BFP16 · ~72 tok/sFP16 · ~75 tok/s
Gemma 3 4BFP16 · ~67 tok/sFP16 · ~70 tok/s
Mistral 7B v0.3Q8_0 · ~75 tok/sFP16 · ~42 tok/s
Qwen2.5 7BQ8_0 · ~71 tok/sFP16 · ~40 tok/s
Qwen2.5 Coder 7BQ8_0 · ~71 tok/sFP16 · ~40 tok/s
DeepSeek R1 Distill Qwen 7BQ8_0 · ~71 tok/sFP16 · ~40 tok/s
Llama 3.1 8BQ8_0 · ~68 tok/sFP16 · ~38 tok/s
Granite 3.3 8BQ8_0 · ~66 tok/sFP16 · ~37 tok/s
Qwen3 8BQ8_0 · ~66 tok/sFP16 · ~37 tok/s
DeepSeek R1 0528 Qwen3 8BQ8_0 · ~66 tok/sFP16 · ~37 tok/s
Gemma 2 9BQ8_0 · ~59 tok/sFP16 · ~33 tok/s
Gemma 4 12BQ8_0 · ~45 tok/sQ8_0 · ~48 tok/s
Mistral Nemo 12BQ8_0 · ~44 tok/sQ8_0 · ~47 tok/s
Gemma 3 12BQ8_0 · ~44 tok/sQ8_0 · ~47 tok/s
Qwen2.5 14BQ6_K · ~48 tok/sQ8_0 · ~39 tok/s
Phi-4 14BQ6_K · ~48 tok/sQ8_0 · ~39 tok/s
Phi-4 Reasoning 14BQ6_K · ~48 tok/sQ8_0 · ~39 tok/s
Qwen3 14BQ6_K · ~47 tok/sQ8_0 · ~38 tok/s
Qwen2.5 Coder 14BQ6_K · ~47 tok/sQ8_0 · ~38 tok/s
DeepSeek R1 Distill Qwen 14BQ6_K · ~47 tok/sQ8_0 · ~38 tok/s
gpt-oss 20BQ5_K_M · ~225 tok/sQ8_0 · ~158 tok/s
Mistral Small 24BQ3_K_M · ~50 tok/sQ6_K · ~31 tok/s
Devstral Small 24BQ3_K_M · ~50 tok/sQ6_K · ~31 tok/s
Magistral Small 24BQ3_K_M · ~50 tok/sQ6_K · ~31 tok/s
Mistral Small 3.2 24BQ3_K_M · ~49 tok/sQ6_K · ~31 tok/s
Gemma 4 26B-A4BQ3_K_M · ~310 tok/sQ6_K · ~194 tok/s
Gemma 2 27BQ2_K · ~56 tok/sQ5_K_M · ~31 tok/s
Gemma 3 27BQ3_K_M · ~43 tok/sQ5_K_M · ~31 tok/s
Qwen3.8 27BQ3_K_M · ~42 tok/sQ6_K · ~27 tok/s
Qwen3.6 27BQ3_K_M · ~42 tok/sQ6_K · ~27 tok/s
Qwen3 30B-A3BQ2_K · ~465 tok/sQ5_K_M · ~258 tok/s
Qwen3 Coder 30B-A3BQ2_K · ~465 tok/sQ5_K_M · ~258 tok/s
GLM-4.7 Flash 30B-A3BQ2_K · ~512 tok/sQ5_K_M · ~283 tok/s
Gemma 4 31BQ2_K · ~49 tok/sQ4_K_M · ~32 tok/s
Nemotron 3 Nano 30B-A3BQ2_K · ~439 tok/sQ5_K_M · ~243 tok/s
Qwen2.5 32BQ2_K · ~47 tok/sQ4_K_M · ~30 tok/s
DeepSeek R1 Distill Qwen 32BQ2_K · ~47 tok/sQ4_K_M · ~30 tok/s
Qwen3 32BQ2_K · ~47 tok/sQ4_K_M · ~30 tok/s
QwQ 32BQ2_K · ~47 tok/sQ4_K_M · ~30 tok/s
Qwen2.5 Coder 32BQ2_K · ~47 tok/sQ4_K_M · ~30 tok/s
Qwen3.6 35B-A3BQ2_K · ~512 tok/sQ4_K_M · ~333 tok/s
Mixtral 8x7BToo bigQ2_K · ~125 tok/s
Llama 3.1 70BToo bigToo big
Llama 3.3 70BToo bigToo big
DeepSeek R1 Distill Llama 70BToo bigToo big
Qwen2.5 72BToo bigToo big
Qwen3 Coder Next 80B-A3BToo bigToo big
GLM-4.5-Air 106B-A12BToo bigToo big
Llama 4 Scout 17B-16EToo bigToo big
gpt-oss 120BToo bigToo big
Qwen3.8 Flash NextToo bigToo big
Qwen3 235B-A22BToo bigToo big
DeepSeek V4 Flash 284B-A13BToo bigToo big
GLM-5.3 Flash 320B-A18BToo bigToo big
MiniMax M3 428B-A23BToo bigToo big
GLM-5 744B-A40BToo bigToo big
GLM-5.3Too bigToo big
DeepSeek V4.1 FlashToo bigToo big
Mistral Large 4 1T-A52BToo bigToo big
DeepSeek V4 Pro 1.6T-A49BToo bigToo big
Qwen3.8 2.4T-A95BToo bigToo big
Kimi K3.1 Pro 2.8T-A104BToo bigToo big