Fast language models at affordable prices.
Our mission is to provide capable models to everyone. We offer a reliable API where you can access Qwen 3.8 27B and other models while only paying for the tokens you use.
## Models
Prices are in USD per million tokens.
| model | input | cached | output | speed |
|---|---|---|---|---|
| qwen-3.8-27b | $0.15 | $0.05 | $1.25 | 80 tok/s |
| deepseek-v4.1-flash | coming soon | |||
| glm-5.3-flash | coming soon | |||
| mimo-v2.5 | coming soon | |||
## Integration
Use Mistbeam with the tools you already have. Connect Claude Code, Codex, Pi, OpenClaw, Hermes Agent, and more with a single config change.
view setup instructions →