Access DeepSeek, Qwen, Kimi, GLM, MiMo, MiniMax and more through one OpenAI-compatible endpoint. Built for engineering teams and enterprise workloads worldwide.
Swap the base URL and your key — the rest of your code stays the same. Streaming, system prompts and JSON output all behave exactly as you'd expect. Change the model string to switch providers.
curl https://token.dabanzi.ai/v1/chat/completions \
-H "Authorization: Bearer sk-tr-..." \
-H "Content-Type: application/json" \
-d '{
"model": "dabanzi/deepseek-v4-flash",
"messages": [{"role": "user", "content": "Explain MoE in one line."}]
}'
# Switch models by changing one string:
# "model": "dabanzi/deepseek-v4-pro"
# More models coming online — full list at /token/modelsSign up with your work email. Your key is ready instantly.
Bank transfer, Alipay or WeChat corporate payment — with an invoice. Your code arrives within one business day of settlement; redeeming credits instantly.
One endpoint, one key. Switch between models by changing a single string.
dabanzi/deepseek-v4-flashdabanzi/deepseek-v4-prodabanzi/qwen3.7-maxdabanzi/qwen3.7-plusdabanzi/kimi-k2.6dabanzi/kimi-k2.7-codedabanzi/kimi-k2.5dabanzi/glm-5.2dabanzi/glm-4.7-flashxdabanzi/minimax-m3dabanzi/mimo-v2.5-prodabanzi/mimo-v2.5openai/gpt-5openai/gpt-5-miniopenai/gpt-4oxai/grok-4mistralai/mistral-largemeta-llama/llama-3.3-70bqwen/qwen3-coderdabanzi/seedream-prodabanzi/seedream-ultradabanzi/seedream-liteNo subscription, no monthly minimum. Per-token billing, priced in USD.
Top up any amount. Pay only for the tokens you use — nothing else.