Supported Models

Poixe AI aggregates leading large models from multiple providers and supports several API protocols, so you can choose the right model and cost for each scenario.

​

Supported model families

We currently focus on text chat models, covering the common model families on the market. Models for image / video / speech generation are being integrated and will be released gradually. Model families covered:

• ChatGPT
• DeepSeek
• Claude
• Gemini
• Grok
• Doubao
• Qwen
• Kimi

New models and versions are added to these families continuously. Refer to the live list on the pricing page for the latest availability.

​

Supported model providers

Poixe AI serves resources through multiple model providers to ensure diversity and stability of service, including (but not limited to):

• OpenAI / Azure OpenAI
• DeepSeek / Anthropic / Google / xAI
• Volcengine / Aliyun / SiliconFlow
• Chutes / Moonshot / Cerebras / Groq
• Fireworks / Novita / Together / Nebius

Providers differ in available models, throughput, stability, and billing policies. We perform basic availability and load management in the background to help you make calls more reliably.

​

Supported API protocols

To work with different ecosystems and SDKs, we support several mainstream protocols:

• OpenAI: Completions, Responses
• Anthropic: Messages
• Google Gemini: generateContent (Gemini Developer API)

If you already have working OpenAI/Anthropic/Gemini code, you usually only need to change base_url and api_key to get started.

​

View the full model list and pricing

For the supported models, per-model pricing, and rate limits, refer to the pricing page: https://poixe.com/pricing

​

Check model status

In the dashboard you can view each model’s health statistics, capacity, and load status: https://poixe.com/model

​

Free models

We offer some free models for testing and light usage: add the :free suffix to the model name to call them. For details, see: https://poixe.com/products/free