Open-weight models for your coding agents
One API for GLM, Kimi and DeepSeek. Works with Claude Code, OpenClaw, Hermes Agent, Cline, Kilo Code and anything OpenAI- or Anthropic-compatible.
Drop-in compatible
OpenAI Chat Completions and Anthropic Messages endpoints. Change the base URL and keep your tools.
Built for long agent loops
Prompt caching with cached tokens billed at the cached rate, and timeouts sized for 300k-token contexts.
Zero data retention
We never store your prompts or completions. Only metadata needed for billing.
Simple prepaid credits
Top up in USD or AUD and pay only for the tokens you use.
Models at launch
- GLM-5.3
- GLM-5.3 Flash
- Kimi K3.1 Pro
- DeepSeek V4.1 Flash
Pay per token with prepaid credits in USD or AUD. No subscription. Pricing is published at launch.
Works with
- Claude Code
- OpenClaw
- Hermes Agent
- Kilo Code
- Cline
- Roo Code
- OpenCode
- Qwen Code
- Cursor
and any OpenAI- or Anthropic-compatible tool
Get in touch
Want early access or have a question? Email us and we will reply as soon as we can.
Email support@openweightrelay.com