DeepSeek text
Useful for low-cost chat, coding assistance, reasoning tests and OpenAI-compatible SDK validation.
Model catalog
Chinese AI models through one OpenAI-compatible gateway.
Start with DeepSeek text models today, request Qwen, Kimi and GLM coverage as your workload expands, and plan early-access media workflows for Kling, Seedance and related Chinese video models.
OpenAI-compatible route
Base URL
https://api.blackeaglecambodia.com/v1AuthBearer token from console
Text endpoint
/chat/completionsModel fieldChoose an available model id
Available and requested models
Start narrow, then expand by workload.
Black Eagle AI is intentionally conservative about availability. The public catalog separates production-ready text access from request-based text families and media-model early access.
DeepSeek's current production naming is
deepseek-v4-flash and deepseek-v4-pro. The older aliases deepseek-chat and deepseek-reasoner are scheduled for retirement on July 24, 2026, so new integrations should use the V4 names now.Qwen family
Good candidate for coding, multilingual apps, agent tools and Chinese-English product workflows.
Kimi / GLM / MiniMax
Useful when the workload needs long context, structured technical output or workflow automation.
Kling video
Text-to-video and image-to-video workflows for creative tools, agencies and product prototypes.
Seedance video
Short-form video generation workflows for ad creatives, social content and automated pipelines.
Happyhours and media models
Additional Chinese media model routes can be evaluated when a team has a specific integration brief.
Integration patterns
What changes in your app?
Most teams start by testing text requests through the same OpenAI-style client they already use.
| Need | Black Eagle setup | What to test first |
|---|---|---|
| Lower-cost chat Support bots, internal tools, content operations. |
Use deepseek-v4-flash through /v1/chat/completions. |
20-50 real prompts, output quality, retry rate and average cost. |
| Reasoning or coding Agents, repository Q&A, technical assistants. |
Start with deepseek-v4-pro, then compare requested Qwen/Kimi/GLM routes. |
Framework compatibility, model id passthrough, streaming and tool-call behavior. |
| Chinese model comparison Need a benchmark across several domestic model families. |
Request Qwen, Kimi, GLM or MiniMax access for a defined workload. | Accepted-answer cost, latency, output style and billing predictability. |
| Video generation Creative API, ads, automated short video workflows. |
Request early-access media onboarding for Kling, Seedance or related providers. | Prompt format, duration, resolution, queue time, webhook or polling needs. |
Tip: if you already have direct Chinese provider accounts, start direct. Black Eagle AI is most useful when global payment, English setup, one key and one credit balance save time. For Qwen, Kimi, GLM, Kling or Seedance, send a model access request. For video workflows, read the Kling and Seedance planning guide. For trust and billing questions, read the FAQ.
Why a gateway
Choose the access model that matches your team.
Black Eagle AI is not the only way to use these models. It is the managed, prepaid path for teams that want to test quickly before handling multiple provider accounts.
Direct provider
Best when you already have Chinese payment, local verification and provider-specific docs handled.
Global router
Best when broad Western model coverage matters more than China-native payment and media access.
Self-hosted gateway
Best for teams with DevOps resources that want to operate their own routing, logs and budgets.
Black Eagle AI
Best for global developers who want prepaid access to Chinese model families through one OpenAI-compatible surface.
Start testing
Open the console, create a token, run one real workload.
Use a small credit package first. If you need Qwen, Kimi, GLM, Kling or Seedance access, describe the workload before topping up larger amounts.