Qwen3.8 27B scores 52 on Artificial Analysis

	Qwen3.8 27B scores 52 on Artificial Analysis(artificialanalysis.ai)
	372 points by anana_ 57 days ago \| 175 comments
	tl;dr: Summary not available.
	HN Discussion: ↑Impressed by benchmark scores showing small model beats much larger frontier models ↑Real-world testing confirms the benchmark numbers reflect genuine capability ↑Questions the economics of massive data centers given efficient small models •Concerns about pricing and inference speed despite the small model size •Reflects on theoretical implications like the lottery ticket hypothesis for shrinking models