Skip to content

Local AI benchmarks by model

Choose a model to compare benchmark scores, generation times and output samples across tested local AI hardware.

Qwen3.8-27B

Language

Compare instruction-following and reasoning scores, plus inference speed across tested hardware.

Qwen3.8-Flash-Next

Language

Compare inference speed across tested hardware and instruction-following scores across quantizations on Olares One.

Qwen-Image-2.1

Image

Compare image generation and editing times, with image samples from Base and Turbo.

MiniMax H3

Video

Compare MiniMax H3 and FastH3 generation times, with video samples at 480p and 720p.