Knowledge base

Tìm nhanh trong tài liệu Zora

Models

Zora currently exposes the Zora GPT/Codex model catalog, Zora image routes, and Zora audio routes.

Programmatic list

Use GET /v1/models with your Zora API key to retrieve the live model list. Looking for effort values? Jump to Reasoning Effort.

How to Choose a Mode

Zora exposes modes as explicit model IDs. Choose the mode by setting the requestmodel field to one of the IDs below. Do not append extra suffixes or private provider names.

Mode	Model ID	When to use it
Default / free smoke test	`gpt-5.5`	Start here for the first API test, free-tier validation, and general chat.
Fast GPT route	`gpt-5.4-mini`	Use when latency matters more than maximum depth.
Quality GPT route	`gpt-5.4`	Use for heavier chat, analysis, and coding prompts.
General GPT route	`gpt-5.2`	Use for balanced general assistant work.
Max coding route	`gpt-5.1-codex-max`	Use for harder repository edits, code review, and agentic coding tasks.
Fast coding route	`gpt-5.1-codex-mini`	Use for quick code edits, explanations, and lightweight agent tasks.
Spark coding route	`gpt-5.3-codex-spark`	Use for small coding tasks and quick agent responses.
Auto review route	`codex-auto-review`	Use for code-review style workflows when available on the account.

Mode selection

{ "model": "gpt-5.4-mini", "messages": [{ "role": "user", "content": "Reply with one short sentence." }] }

Reasoning Effort

For text models, set effort with a request body parameter. Keep the model field as an exact model ID, then add effort or reasoning_effort. Supported values are:

Effort	When to use it
`low`	Fastest setting. Use for simple prompts, smoke tests, and short answers.
`medium`	Balanced setting. Use for normal coding, explanations, and support tasks.
`high`	Deeper setting. Use for harder debugging, planning, and multi-step analysis.
`xhigh`	Most careful setting. Use for long coding tasks, reviews, and critical checks.

Chat Completions

{ "model": "gpt-5.5", "effort": "high", "messages": [{ "role": "user", "content": "Debug this issue step by step." }] }

Responses

{ "model": "gpt-5.4-mini", "reasoning_effort": "medium", "input": "Summarize this repository change." }

Do not encode effort in the model name. For example, use model: "gpt-5.5" pluseffort: "high", not a suffixed model ID.

Zora GPT/Codex Models

Use the exact model IDs returned by GET /v1/models. Production currently routes these text/Codex IDs through the internal model router.

Model	Context	Endpoint	Access
`gpt-5.5`	Zora	`/v1/responses` `/v1/chat/completions`	Free
`gpt-5.4`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.4-mini`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.3-codex`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.2-codex`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.2`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.1-codex-max`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.1-codex-mini`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5-codex`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5-codex-mini`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`gpt-5.3-codex-spark`	Zora	`/v1/responses` `/v1/chat/completions`	Paid
`codex-auto-review`	Zora	`/v1/responses` `/v1/chat/completions`	Paid

Legacy-Compatible Routes

The current public model catalog does not advertise legacy aliases. Use the exact Zora model IDs from GET /v1/models.

Model	Endpoints	Access

Zora Audio Routes

Audio routes are available on paid plans and PAYG credit for async speech, transcription, dubbing, sound effects, and voice tools.

Model	Endpoint	Access
`whisper-large-v3`	`/v1/audio/transcriptions` `/v1/audio/translations`	Paid
`whisper-large-v3-turbo`	`/v1/audio/transcriptions`	Paid
`eleven_multilingual_v2`	`/v1/audio/speech`	Paid
`eleven_v3`	`/v1/audio/speech`	Paid
`eleven_flash_v2_5`	`/v1/audio/speech`	Paid
`eleven_turbo_v2_5`	`/v1/audio/speech`	Paid
`eleven_multilingual_sts_v2`	`/v1/audio/voice-changer`	Paid
`eleven_text_to_sound_v2`	`/v1/audio/sound-effects`	Paid

Zora Image Routes

Image routes are available on paid image-enabled plans and image-only monthly keys.

Model	Endpoint	Access
`gpt-image-2`	`/v1/models`	Paid

Free Tier

Free accounts can call only these model IDs:

Free model IDs

gpt-5.5