Intel Arc B580 vs. AMD Radeon RX 9000 vs. NVIDIA RTX 50 Series For Llama.cpp Vulkan Performance

Written by Michael Larabel in Display Drivers on 8 December 2025 at 10:50 AM EST. Page 4 of 5. 11 Comments.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: gpt-oss-20b-Q8_0, Test: Text Generation 128. RTX 5090 was the fastest.

With GPT-OSS 20B the RX 9070 series came out slightly ahead of the RTX 5070 in the TG128 benchmark. The Radeon RX 9070 performance-per-Watt here was also comparable to the NVIDIA Blackwell GPUs. The Radeon RX 9000 series were delivering nice value here.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: gpt-oss-20b-Q8_0, Test: Prompt Processing 512. RTX 5090 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: gpt-oss-20b-Q8_0, Test: Prompt Processing 1024. RTX 5090 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: gpt-oss-20b-Q8_0, Test: Prompt Processing 2048. RTX 5090 was the fastest.

For the prompt processing benchmarks with GPT-OSS 20B, the Radeon RX 9070 series continued performing decent relative to the RTX 5060 Ti and the RTX 5070 12GB not being able to run for lack of vRAM.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: DeepSeek-R1-Distill-Llama-8B-Q8_0, Test: Text Generation 128. RTX 5090 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: DeepSeek-R1-Distill-Llama-8B-Q8_0, Test: Prompt Processing 512. RTX 5090 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: DeepSeek-R1-Distill-Llama-8B-Q8_0, Test: Prompt Processing 1024. RTX 5090 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: Vulkan, Model: DeepSeek-R1-Distill-Llama-8B-Q8_0, Test: Prompt Processing 2048. RTX 5090 was the fastest.

For DeepSeek-R1-Distill-Llama-8B with the Llama.cpp Vulkan back-end, the NVIDIA GeForce RTX 50 series graphics cards tested continue delivering much stronger performance.

11 Comments - Next Page

Page: 1 2 3 4 5 Next Page

URL: https://www.phoronix.com/review/llama-cpp-vulkan-eoy2025/4

⇱ Intel Arc B580 vs. AMD Radeon RX 9000 vs. NVIDIA RTX 50 Series For Llama.cpp Vulkan Performance - Phoronix

Intel Arc B580 vs. AMD Radeon RX 9000 vs. NVIDIA RTX 50 Series For Llama.cpp Vulkan Performance