FAQ
Consultez les bases sur l'accès aux modèles, les tarifs et l'utilisation de l'API.
Langue des données en direct
Le contenu complet en direct — dates, classements et sources externes — est affiché ci-dessous dans sa langue source, l’anglais. Les éléments non traduits conservent le libellé d’origine afin de privilégier l’exactitude des mises à jour.
Coverage can include GLM, Qwen, DeepSeek, Kimi, and MiniMax for LLM workloads, Qwen Image and Wan for image generation, and Seedance, Hailuo, and Kling for video generation, depending on availability, use case, and qualified usage volume.
No. The first buyer is often a business owner, agency lead, product leader, marketing team, AI application company, or system integrator that needs lower-cost model access or creative production capacity.
Every account can call every model. Accounts start on Standard pricing after sign-up. One top-up of any amount moves the whole account to Paid pricing: models with a Paid price bill at it, and the rest stay at Standard. For larger volumes, contact ChinaAPI about Enterprise pricing.
Yes. Every model page and the dashboard model catalog list the Standard price in USD, and the Paid price where a model has one. Standard model rates follow providers' international list prices at 100%; some models have a separate Paid rate after the first top-up.
Yes. Pilot access is designed for small-scope evaluation before expanding into volume-based pricing.
Yes. Add credits by bank card in Dashboard Wallet where offered. Cryptocurrency and local wallet methods may also appear for your account. Redemption codes remain an option for offline purchases. The dashboard shows current methods and amounts at checkout.
No. ChinaAPI provisions access for international teams. Integration happens over standard APIs from your existing infrastructure, with no local entity required.
Often, for the right workloads. The savings come from routing suitable tasks to lower-cost model families rather than replacing every call at once, so each workload should be evaluated on its own.
Yes. Pilots can be structured to compare model families side by side on the same workload, and the integration path is OpenAI-compatible, so switching models does not require rebuilding your stack.
For LLMs: RAG, customer support automation, SaaS inference, and internal tools. For creative teams: product videos, ad variants, social content concepts, and high-volume image or video iteration.
Company, country, current model stack, monthly AI spend, expected volume, target use case, preferred model families, and whether email, WhatsApp, or Telegram is best for follow-up.
