SWT-API provides access to various AI models including Claude, Kimi, GLM, and Gemini for chat and completion tasks.

SWT-API is a self-use platform offering API access to multiple AI models. Key models include Claude-Sonnet-4.5-Poe, Kimi-K2-Thinking, GLM-4.6-Cerebras/GLM-4.6-Cs-Pay, Gemini-3-Pro-Poe/Gemini-3-Pro-Preview, and Gemini-2.5-Pro/-Preview-06-05. The service operates through multiple endpoints: a primary site (api.lhyb.dpdns.org), a secondary site (api2.lhyb.dpdns.org), and a backup site (new-api.koyeb.app). It supports typical AI use cases such as chat and text generation. Note: Poe's GPT5 and 5.1 series do not support streaming calls, and paid models do not guarantee availability.

Time	Model	Speed	Latency
Sep 12, 09:52 AM	sonoma-sky-alpha	139.35 t/s	5.79s
Sep 12, 09:50 AM	gemini-2.5-flash	139.35 t/s	9.49s
Sep 12, 09:40 AM	gpt-5-mini-poe	88.90 t/s	6.96s
Sep 12, 09:30 AM	gpt-5-chat-poe	64.03 t/s	1.72s
Sep 7, 03:02 AM	gemini-2.5-flash-nothinking	131.60 t/s	12.98s
Sep 6, 07:00 AM	sonoma-sky-alpha	128.11 t/s	3.27s
Sep 5, 03:07 AM	gpt-5-poixe	52.58 t/s	17.07s
Sep 5, 03:06 AM	gpt-5-chat-poixe	73.29 t/s	1.75s
Sep 5, 02:59 AM	grok-4-poixe	37.06 t/s	6.64s
Sep 5, 02:57 AM	gpt-5-poe	65.96 t/s	2.24s

SWT-API

Supported Models

Recent Test Records

SWT-API

About SWT-API

Supported Models

Recent Test Records