Global service map
. Regenerated with each Chinaready Landscape publish.
Quick answer: Anthropic blocks all model access from mainland China IPs — applications and users in mainland China cannot call overseas Anthropic Claude model APIs. Use mainland China LLM services instead. Most of those providers are compatible with both the OpenAI and Anthropic API shapes — so your application may keep working with little or no code change. Chinaready currently maps Anthropic to DeepSeek, Moonshot AI (Kimi), Qwen (Alibaba Cloud Model Studio). Availability in China: Unavailable.
Availability in China: Unavailable
Also searched as: anthropic claude, claude, claude api.
Anthropic blocks all model access from mainland China IP addresses. If the application or the calling client runs in mainland China, overseas Anthropic Claude model endpoints are not a usable production dependency. Plan on mainland-reachable LLM APIs instead of assuming a global Anthropic key will work from China infrastructure.
The practical upside: China foundation-model providers listed below almost all expose APIs that are compatible with OpenAI and Anthropic client conventions. Many teams can point existing SDKs or HTTP clients at a China endpoint (base URL, key, and model name) without rewriting the application.
Use the mapped candidates below as a research shortlist, then confirm API compatibility, model behavior, and compliance for your stack.
DeepSeek provides open-weight and API-accessible foundation models optimized for reasoning and coding with competitive inference cost.
China context: DeepSeek offers mainland-reachable APIs and widely used open weights as a practical alternative when OpenAI or Anthropic are unavailable for China production stacks.
Moonshot AI provides Kimi long-context language models and APIs for chat, document analysis, and application development.
China context: Kimi provides mainland-reachable long-context model APIs for teams replacing OpenAI or Anthropic in document-heavy China applications.
Alibaba Cloud Model Studio provides managed access to Qwen models and DashScope APIs for building generative AI applications.
China context: Model Studio and DashScope provide mainland-hosted Qwen APIs and enterprise cloud integration for China deployments that cannot rely on OpenAI, Anthropic, or other global model endpoints.
Zhipu AI provides bilingual GLM foundation models and APIs for enterprise chat, reasoning, and application development.
China context: Zhipu AI offers mainland-reachable bilingual model APIs and enterprise services for production workloads that need a China-based alternative to OpenAI or Anthropic.
MiniMax provides multimodal foundation models and developer APIs for text, speech, image, video, and agent applications.
China context: MiniMax offers mainland-reachable multimodal APIs for China applications that need local text, speech, image, or video generation instead of OpenAI or Anthropic services.
Volcengine Ark provides managed access to ByteDance Doubao models and model services for enterprise generative AI development.
China context: Ark provides mainland-hosted Doubao model endpoints and Volcengine integration for China workloads that cannot rely on OpenAI or Anthropic, though model behavior and platform interfaces are not fully drop-in compatible with global APIs.
No. Anthropic blocks all model access from mainland China IPs, so applications and users in mainland China cannot call overseas Anthropic Claude model services. Chinaready labels Anthropic as Unavailable for mainland China production stacks. Use a mainland-reachable LLM API instead.
Usually yes. Most mainland China foundation-model providers expose APIs compatible with OpenAI and Anthropic client shapes. Many applications can keep working by changing the base URL, API key, and model name rather than rewriting application logic.
Chinaready Landscape currently lists these China-market options for Anthropic: DeepSeek, Moonshot AI (Kimi), Qwen (Alibaba Cloud Model Studio), Zhipu AI (GLM), MiniMax, Volcengine Ark (Doubao). Replacement fit varies by product, so treat this as a research shortlist rather than a one-to-one endorsement. Prefer providers whose OpenAI- or Anthropic-compatible endpoints match your existing client stack.
Often closer than for other SaaS categories: many China LLM APIs are OpenAI- and Anthropic-compatible, so the migration is frequently configuration rather than a full rewrite. Still validate model quality, rate limits, compliance, and feature gaps before production cutover.
Use the interactive Chinaready Landscape to compare adjacent AI model and platform services, then read Chinaready's main site for launch operating guidance covering compliance, distribution, and go-to-market constraints beyond vendor selection. If the path remains unclear, book a call with Chinaready.
