A unified API gateway for large language models, offering competitive pricing and stability.

[503]gemini Flash Lite[503]gemini FlashGemini 2.5 Pro[503]gemini ProClaude 4 SonnetDeepSeek V3DeepSeek R1[free]deepseek V3

鲨鱼魔法 provides a unified API gateway for accessing large language models. Users can integrate by replacing the model base URL with provided endpoints like '酒馆 URL', '完整（鬼谷） URL', or '域名 URL'. The platform emphasizes better pricing and stability. It includes a model marketplace and supports QQ email for account registration. Key features include a console for API key management, documentation, and a QQ group for customer support. The service is intended for legal learning and research purposes, with compliance measures for AI-generated content. Weekly maintenance occurs on Saturday or Sunday mornings.

Time	Model	Speed	Latency
Aug 12, 02:51 PM	[free]deepseek-v3	0.00 t/s	-s
Aug 12, 02:48 PM	grok-3	28.05 t/s	6.44s
Aug 12, 02:47 PM	claude-4-sonnet	50.48 t/s	3.47s
Aug 12, 02:37 PM	deepseek-r1	12.80 t/s	42.03s
Aug 12, 02:32 PM	deepseek-v3	22.97 t/s	2.89s
Aug 12, 02:26 PM	gemini-2.5-pro	86.07 t/s	53.92s
Aug 12, 02:25 PM	[503]gemini-flash-lite	229.64 t/s	2.17s
Aug 12, 02:23 PM	[503]gemini-flash	199.70 t/s	10.04s
Aug 12, 02:20 PM	[503]gemini-pro	66.39 t/s	20.63s

鲨鱼魔法

Supported Models

Recent Test Records

鲨鱼魔法

About 鲨鱼魔法

Supported Models

Recent Test Records