# ABot-World-0-5B API

> ABot-World-0-5B as a hosted realtime API. A third-person navigable world stream on main_video. $0.0023 per billable second.

Source: https://windflow.online/models/abot-world-0-5b · Navigable world model · Windflow Labs

## What it is

ABot-World is an open-source interactive world model from Alibaba's AMAP computer-vision lab, aimed at rolling out an endless world in real time. ABot-World-0-5B is served on Windflow as a third-person, keyboard-driven world: movement drives an on-screen character rather than a free camera.

## Specifications

- **Architecture:** 5B keyboard-conditioned world model (Wan2.2-TI2V backbone, LongForcing causal student)
- **Resolution:** 1344 × 768 delivered (896 × 512 generation + learned super-resolution)
- **Chunk:** 12-frame AR blocks; held keys are sampled at each block boundary
- **Inputs:** A seed image, a text prompt, and persistent movement/camera key state
- **Outputs:** A third-person navigable world stream on main_video
- **Delivery:** WebSocket (H.264)
- **Session protocol:** sdk.realtime.v2
- **Price:** $0.0023 per billable second

## Pricing

$0.0023 per billable session-second ($8.28/hr of billable time). Billing starts at the session's first generation_started event and stops when the session is released or ends. Queueing, worker startup, time before generation starts and the documented reconnect grace are not billed. After generation starts, paused or unwatched time remains billable; release the owner session to stop the meter.

## Use cases

- Third-person explorable scenes with a controllable character.
- Game-style prototypes driven by an eight-key control vocabulary.
- Endless-rollout experiments where the world keeps generating as the character moves.
- Comparing third-person and first-person world models on one protocol.

## Where it comes from

- [amap-cvlab/ABot-World on GitHub](https://github.com/amap-cvlab/ABot-World) — Apache-2.0

## FAQ

### What controls does ABot-World use?

The model's eight-key vocabulary: WASD to move and IJKL for the camera, quantised to the next generation block.

### What resolution is delivered?

1344 × 768, produced by learned super-resolution over an 896 × 512 generation canvas.

### How much does ABot-World-0-5B cost on Windflow?

ABot-World-0-5B costs $0.0023 per billable session-second ($8.28/hr of billable time). Billing starts at the session's first generation_started event and stops when the session is released or ends. Queueing, worker startup, time before generation starts and the documented reconnect grace are not billed. After generation starts, paused or unwatched time remains billable; release the owner session to stop the meter. There is no subscription.

### How do I get access to ABot-World-0-5B?

Windflow is in private beta. Request access from the sign-up page, or ask for a live demo of ABot-World-0-5B. Approved accounts start with $5 and four concurrent sessions.

### Is ABot-World-0-5B always running?

No. Model capacity is brought up on demand rather than kept standing. Live readiness for each model is published at api.windflow.online/v1/status, and a live demo can be arranged on request.

## Related guides

- [Choosing a realtime video model: a decision guide for 12 hosted models](https://windflow.online/guides/choosing-a-realtime-video-model)
- [World model API: embed a navigable AI world in your app](https://windflow.online/guides/world-model-api)
- [World model seed images and prompts: a practical guide](https://windflow.online/guides/world-model-seed-images-and-prompts)
- [What is a world model? How it differs from video generation](https://windflow.online/guides/what-is-a-world-model)
- [How to recover a realtime video stream after a dropped connection](https://windflow.online/guides/recover-a-dropped-realtime-video-stream)
- [How to use the ABot World API](https://windflow.online/guides/abot-world-api)