计费说明
0 条价格记录Hugging Face acts as a router: its official pricing docs state it adds no markup and that you pay the underlying provider, while requests served by its own 'hf-inference' provider are charged for compute time multiplied by the price of the underlying hardware (official example: a job taking 10 seconds on a GPU that costs $0.00012 per second is billed $0.0012). Platform plans are subscriptions (PRO $9/month; Team $20 per user/month; Enterprise $50 per user/month; free users get monthly inference credits). Because billing is compute-time / per-provider rather than a single unified unit price, no unified USD unit price is recorded.
价格明细
不同模态使用真实计费单位暂无统一公开价格
官方价格页存在，但没有统一单价；请以官方页面为准。
来源与核验
legacy-unreviewed迁移保留旧内容；来源是旧链接线索，本次未重新打开核验。