Hub Open Source
Open-weight reasoning models often use far more tokens than closed models, making them less efficient per query, according to Nous Research. Models like DeepSeek and Qwen use 1.5 to 4 times more tokens than OpenAI and Grok-4—and up to 10 times more for simple knowledge tasks. Mistral's Magistral models stand out for especially high token use. 
Average tokens used per task by different AI models. | Image: Nous ResearchIn contrast, OpenAI's gpt-oss-120b, with very short reasoning paths, shows that open models can be efficient, especially for math problems. Token usage depends heavily on the type of task. Full details and charts are available at Nous Research.
Wan2.2 A14B now tops the rankings for open source video models, according to Artificial Analysis. It ranks seventh for text-to-video and fourteenth for image-to-video, with the lower placement in the latter likely due to its 16 frames per second output compared to 24 fps in some competitors. Among open models, Wan2.2 A14B leads the field, but it still trails behind closed models like Veo 3 and Seedance 1.0 in overall performance. Pricing, however, is often much lower depending on the provider.
