One key.Open the door to AI.
Our own computing infrastructure and official channels — more choice, reliable access, lower costs, with data kept in China. Ready for Claude Code and Codex.
Our own infrastructure. More than aggregation.
We run our own infrastructure alongside upstream access. Leading Chinese open models run on our accelerator clusters, optimized from inference kernels to scheduling. Owning the compute gives us control over availability, latency and pricing. Official Chinese provider channels run in parallel, with automatic failover when a route is disrupted.
OpenAI and Anthropic compatibility
Both OpenAI and Anthropic protocols are supported. Migrate existing code by changing the base URL, without replacing your SDK.
from openai import OpenAI
client = OpenAI(
base_url="https://api.goodputai.cn/v1",
api_key="GOODPUT_API_KEY",
)
resp = client.chat.completions.create(
model="deepseek/deepseek-v4-pro", # Use any model
messages=[{"role": "user", "content": "Hello"}],
)Built for agent tools
Connect Claude Code, Codex, Cline and Roo Code out of the box, with normalized protocol differences for consistent long-running tool calls, streaming and prompt caching.
# ~/.zshrc or ~/.bashrc
export ANTHROPIC_BASE_URL="https://api.goodputai.cn"
export ANTHROPIC_AUTH_TOKEN="YOUR_GOODPUT_API_KEY"
claude # Restart your terminal, then launch to connectOur own deployments. Flexible pricing.
Leading open models run on our own Chinese accelerators. Control over costs lets us offer flexible pricing, including lower off-peak rates and reductions as our clusters improve.
View model pricingData stays in China
Our clusters run inference in China, and Chinese provider channels keep the processing chain in China. Enterprise authentication and isolated quotas support security and compliance controls.
Availability illustration, not a live metric or SLA commitment
Reliable supply with automatic failover
Our own clusters underpin availability and latency. Official Chinese provider channels provide parallel backup, with automatic failover during rate limits or failures to keep applications running.
Ready for enterprises and teams
Enterprise login, organizations and teams, two-level quota controls and detailed usage reporting give members secure access to Chinese models, with administrators managing costs and permissions.
One organization, multiple teams
Create multiple teams within an organization, with team membership separating resources and permissions.
Two-level quota controls
Combine team governance limits with individual member quotas. Administrators decide how much each member may use.
Detailed usage reporting
Review usage and costs by team, member, model and day.
Consolidated billing
An organization prepaid balance and periodic billing bring everyone onto one bill.
Invoice requests
Request invoices from usage bills to simplify finance work.
Popular models
MiniMax H3 video generation supports text-to-video, image-to-video (first frame or first/last frames), and reference-to-video using text, images, audio and video. Output: 768P, with audio generated by default. Billed by successfully generated duration at a list price of CNY 0.50/second; promotional pricing is shown on the page. Failed generations are not billed. Reference images, audio and video inputs are temporarily free.
MiniMax H3: text/image/reference-to-video (first or first/last frames), with text/image/audio/video inputs. 768P; audio on by default. List CNY 0.50/s; promo price on page. Failures unbilled; reference inputs temporarily free. Fast: pull-mode queue; submit returns task id. Only 1344x768 (16:9, default) or 768x1344 (9:16); 5–14s (default 5s). Frame rounding may add up to ~0.7s; billed by requested seconds. Other parameters match H3.
GLM-5.3 is the next-generation model in Z.AI's GLM series, supporting up to 1M context tokens and 128K output tokens. Native deep thinking, tool calling and structured output support long-document understanding, code generation and complex reasoning. Served on our own compute infrastructure, with both OpenAI and Anthropic API protocols.
Example scenarios
Illustrative scenarios, not verified customer testimonials.
“After switching the default Claude Code model to Kimi K2.7 on self-operated infrastructure, a team cuts its monthly model spending by an order of magnitude, with little difference in task completion.”
“One agent framework runs DeepSeek and GLM by changing model names, without changing business code. Supporting both protocols simplifies integration.”
“A fallback model list keeps a promotion running by switching automatically to a backup model when the primary channel is rate limited, without interrupting users.”
Create an account and start building
Change a base URL and give your team access to leading Chinese models.