Models & Plans

Start with these three values

Every OpenAI-compatible tool setup needs:

• base_url: {{BASE_URL}}
• api_key: your Prysm API key from the console
• model: the exact request model name from model marketplace

What the models page shows

The models page is the source of truth for model selection. Each card includes:

• exact model name
• provider
• pricing
• capability tags

Common tags on the page include:

• Reasoning
• Coding
• Vision
• Tools
• Chat
• Free
• Multimodal

Use search and tags together to narrow the catalog quickly.

Free / Pro / Enterprise / Enterprise Golden

Plans attach to teams. The model list shown in this guide is an example set; the exact usable models come from the models assigned to the team by an administrator.

Plan	Billing	Credit	Seats	Models	RPM	TPM
Free	¥0	¥5 registration credit, one-time	1	Configured Free models	10	50,000
Pro	¥599/month	1,200,000 credits/month	1	Standard model pool, excluding overseas models	30	300,000
Enterprise	Usage-based token billing	Contract configurable	100	Standard model pool, excluding overseas models	300	5,000,000
Enterprise Golden	Usage-based token billing	Contract configurable	100	Standard model pool + overseas models	500	10,000,000

How to choose the right model

1. Choose the team plan that matches your workload.
2. Open model marketplace.
3. Search by model name, provider, price, or capability tags.
4. Copy the exact model name shown on the card.
5. Send one test request in Playground, Claude Code, VS Code, or Cline.

Recommended plan fit

Free

Use Free for:

• initial trial
• learning workflows
• low request volume
• validating a tool connection

Pro

Use Pro for:

• personal paid access
• everyday coding
• lightweight agent workflows
• monthly credit balance

Enterprise

Use Enterprise for:

• organization workloads
• standard model pool access
• higher request volume
• budget and audit workflows

Enterprise Golden

Use Enterprise Golden for:

• overseas model access
• cross-border workloads
• high throughput
• production critical tasks

Model page habits that save time

• Use the exact request model name from the model card.
• Check the provider and pricing before switching tools.
• Use the pricing field to compare cost-sensitive options.
• Use capability tags to separate coding, reasoning, and multimodal models.

Common mistakes

• Using a display label in place of the request model name
• Choosing a model outside the active team model scope
• Sending a valid key with the wrong model name
• Filling a base URL field with {{CHAT_COMPLETIONS_URL}}

Use {{BASE_URL}} for provider root / base URL fields. Use {{CHAT_COMPLETIONS_URL}} only when a tool asks for the full chat completions endpoint.