Skip to main content

Human-readable model names and copyable API IDs are not the same thing.

Use the display name to compare models, then copy the exact API ID from the live authenticated catalog. Availability follows the key’s scope and active production routes.

Format
Reference catalog
Providers
5
Reference IDs
26
Scope
Key-dependent

Three checks before selecting a model.

  1. 01

    Check key scope

    Claude-only, OpenAI-only, combined, API-credit, and exact-model keys can return different catalogs.

  2. 02

    Copy the API ID

    Client configuration uses the exact machine-readable ID, including punctuation and version suffixes.

  3. 03

    Check capabilities

    Tools, images, reasoning controls, context limits, and output limits are model- and route-specific.

Models represented in the shared CloudService catalog.

This reference explains names and IDs. It is not an availability guarantee; the live catalog filters models to currently active routes for the requesting key.

Anthropic

7 reference IDs

Display nameAPI IDDescription
Claude Haiku 4.5claude-haiku-4-5-20251001Fast Anthropic model for low-latency chat and everyday work
Claude Opus 4.6claude-opus-4-6High-capability Anthropic reasoning model for complex work
Claude Opus 4.7claude-opus-4-7High-capability Anthropic model for advanced coding and reasoning
Claude Opus 4.8claude-opus-4-8Premium Anthropic model for deep analysis and demanding tasks
Claude Opus 5claude-opus-5Latest premium Anthropic model for advanced coding, agents, and deep reasoning
Claude Sonnet 5claude-sonnet-5Balanced Anthropic model for strong reasoning and production workflows
Fable 5fable-5Anthropic-family Fable model through CloudService

CloudService

8 reference IDs

Display nameAPI IDDescription
Codex Auto Reviewcodex-auto-reviewCloudService-optimized automated code review and PR analysis
GPT-5.3 Codexgpt-5.3-codexBudget-friendly coding assistant
GPT-5.4gpt-5.4Advanced long-context coding and analysis
GPT-5.4 minigpt-5.4-miniFast low-cost everyday development
GPT-5.5gpt-5.5Premium reasoning and agentic coding
GPT-5.6 Lunagpt-5.6-lunaAvailable through CloudService
GPT-5.6 Solgpt-5.6-solAvailable through CloudService
GPT-5.6 Terragpt-5.6-terraAvailable through CloudService

Gemini

8 reference IDs

Display nameAPI IDDescription
Gemini 2.5 Flashgemini-2.5-flashFast Google Gemini model for low-latency chat and everyday work
Gemini 2.5 Flash Litegemini-2.5-flash-liteLightest Google Gemini model for high-volume, cost-sensitive work
Gemini 3 Flash Previewgemini-3-flash-previewPreview Google Gemini 3 model tuned for fast responses
Gemini 3 Pro Previewgemini-3-pro-previewPreview Google Gemini 3 model for advanced reasoning and long context
Gemini 3.1 Flash Lite Previewgemini-3.1-flash-lite-previewPreview lightweight Google Gemini model for high-volume work
Gemini 3.1 Pro Previewgemini-3.1-pro-previewPreview Google Gemini model for demanding reasoning and agentic work
Gemini 3.5 Flashgemini-3.5-flashFast Google Gemini model for everyday coding and chat
Gemini 3.6 Flashgemini-3.6-flashLatest fast Google Gemini model for everyday coding and chat

Grok

1 reference ID

Display nameAPI IDDescription
Grok 4.5grok-4.5xAI Grok model for reasoning, coding, and general assistance

Qwen

2 reference IDs

Display nameAPI IDDescription
Qwen 3.6 Plusqwen3.6-plusAlibaba Qwen model for general reasoning and coding
Qwen 3.7 Plusqwen3.7-plusLatest Alibaba Qwen model for general reasoning and coding

The authenticated response wins.

Before a production request, fetch the catalog with the same key and base URL the client will use.

Check current models