---
slug: model-specs
lastmod: 2026-06-05
author: ai-image-research-editor
title: AI Pin Maker Model Specs and Capabilities
description: Compare AI image, video, and audio models on AI Pin Maker by capability, resolution, duration, cost, and content tier — sourced from the live router.
keyword: AI Pin Maker model specs
h1: AI Pin Maker Model Specs and Capabilities
summary: Specs and capabilities for every image, video, and audio model available through the AI Pin Maker router.
image: https://oss.axis-ai.dev/oss/new-api-dev/2026/06/05/image/gpt-image-2/channel-1/user-1/task_y5doizcxll4xvhozann4jqqj0yjfceyn.png
imageAlt: AI Pin Maker model specs cover with comparison grid
appPath: /zh/eshi/text-to-image
cta: Open the Studio to Try a Model
articleUrl: /articles/ai-mascot-generator-pin-brand-workflow/
articleLabel: Read a workflow that picks the right model
order: 212
reviewedBy: ai-image-research-editor
reviewedDate: 2026-06-05
---

## User intents
- Pick the right model before spending credits on a generation
- Compare image and video models by resolution, duration, and cost
- See which models support a wider creative tier vs standard work-safe routing
- Decide between same-family versions (newest vs stable vs legacy)
- Confirm a model ID before adding it to a workflow script

## Step-by-step
- Identify the capability you need first (text-to-image, text-to-video, image-to-video, reference-to-video)
- Pick the tier (standard work-safe vs mature fictional) before picking a model
- Compare same-family versions, then pick the version closest to your quality vs cost target
- Run a small dry run before a full paid generation, especially after a model upgrade
- Re-check this page when the changelog flags a model addition or retirement

## FAQ
### How are the AI Pin Maker model specs kept up to date?
The specs are sourced from the live router and refreshed when the changelog logs a model addition, version bump, or pricing change. The exact model IDs match what the [text to image](https://aipinmaker.com/zh/eshi/text-to-image?feature=model-specs) and [text to video](https://aipinmaker.com/zh/eshi/text-to-video?feature=model-specs) studios call.

### Which AI image model should I pick for the best quality?
For broad editorial quality, start with the latest GPT Image route or the Gemini 3 Pro image route in the [text to image](https://aipinmaker.com/zh/eshi/text-to-image?feature=model-specs) studio. For consistent character or scene reuse, compare Wan 2.7 image-pro and current Seedream models.

### Which AI video model is best for short product clips?
For short, dialogue-free product clips, compare Veo 3.0 and Kling v3 in the [text to video](https://aipinmaker.com/zh/eshi/text-to-video?feature=model-specs) studio. For longer or multi-shot work, see the Wan 2.6 family and Seedance 2.0 in the same studio.

### Is using model specs free?
Yes. Browsing the model specs page is free, no sign-up. Credits required: 0. Generating with a specific model on the studio does consume credits.

### What is the difference between Current, Latest, Stable, and Legacy status tags?
Current is the recommended default for that capability. Latest is the newest released version, which may still have edge cases. Stable is a previous version that has matured in production. Legacy means the model is kept available for compatibility but is no longer recommended for new workflows.

### Where can I see the full model ID list?
The router exposes a `/v1/models` endpoint that returns the live model list with capabilities, supported parameters, and pricing estimates. Use it before scripting a workflow.

### Are mature fictional models listed here?
Yes, but they are clearly tagged with the tier so you can filter them out for work-safe creative runs. Tier routing on the [text to image](https://aipinmaker.com/zh/eshi/text-to-image?feature=model-specs) studio respects the tag.

### How should I compare same-family versions?
Use the in-page comparison table that keeps the first two rows (max duration and multi-shot or its image equivalent) constant across versions, and lets the later rows shift to the newest version's headline feature.