The performance gap between leading AI models from U.S. tech companies and the best Chinese open-source models has shrunk to just 4.4 months, according to Mozilla’s recent State of Open Source AI report, reviewed by Ars Technica. The results suggest that for most organizations, it is no longer worth choosing more expensive closed models to perform routine tasks.
The performance gap between leading AI models from U.S. tech companies and the best Chinese open-source models has shrunk to just 4.4 months, according to Mozilla’s recent State of Open Source AI report, reviewed by Ars Technica. The results suggest that for most organizations, it is no longer worth choosing more expensive closed models to perform routine tasks.