Open source model closes gap on proprietary benchmarks
A community-trained release now scores within a few points of the closed-source leaders on two major evaluations.
A newly released open source model has closed much of the performance gap with proprietary leaders, scoring within a few points on two widely cited benchmarks while remaining freely available for commercial use.
The release has been welcomed by developers who cited cost and data control as reasons for preferring open weights, even where a small performance gap remains against the top closed models.
