Official site ↗ · 🇺🇸 United States
Serves: 200+ modelos de pesos abiertos a traves de 18 proveedores: Baseten, Cerebras, Cohere, DeepInfra, Fal AI, Featherless AI, Fireworks, Groq, HF Inference, Novita, Nscale, OVHcloud AI Endpoints, Public AI, Replicate, Scaleway, Together, WaveSpeedAI y Z.ai
Billing unit: tokens (pass-through provider rate)
Converts to tokens: full
Verifiable pre-pay: yes
Pricing / discount: no markup — you pay the partner's published per-token rate; monthly credit included (Free $0.10, PRO $2, Team / Enterprise $2/seat)
A router, not a host: one HF token routes to leading inference providers, OpenAI-compatible (router.huggingface.co/v1), with no HF markup — you pay the partner's rate. The provider is chosen by policy suffix (:fastest default, :cheapest, :preferred, or an explicit :provider). A small monthly credit ($0.10 free / $2 on PRO at $9/mo) then pay-as-you-go. It serves whatever its partners serve (open weights only), so price and quantization come from the chosen partner, not HF. Closed models (GPT-5.x, Claude) are not routed. 2026-09-24: creditos y el 'sin markup' sin cambios. La lista de socios ha crecido a 18, y cuatro de ellos tienen ficha propia en este mismo fichero: Featherless, Nscale, OVHcloud y Public AI. 2026-10-02: sin cambios. Matiz nuevo: al usuario gratuito, agotado su credito de $0,10, el pago por uso le exige COMPRAR creditos antes.