BytePlus ModelArk
BytePlus's hosted model platform: pay-per-token API access to ByteDance's Seed (Dola-Seed) language models and Seedance video, Seedream image and 3D generation models, alongside third-party models such as DeepSeek and GLM, with fine-tuning and a console.
Find alternatives to BytePlus ModelArk
What it does
- Model inference
- Text generation
- Video generation
- Image generation
Sources
- Pricing
-
Per-million-token rates by model, for example seed-2-0-lite at USD 0.25 input and 2.00 output for prompts up to 128K. New users get a one-time trial ("Get started with 500,000 free tokens for every Large Language Model and 2,000,000 free tokens for every vision model."), which is not a recurring free tier.
- Description
-
"Features and unit prices vary with model services. This topic introduces the billing formulas and unit prices for each model"; the price list has language, video, image, 3D and embedding models, including Seed (now branded Dola-Seed), DeepSeek and GLM rows.
- What it does
-
Hosted inference over text, video and image models.
- Sold within
-
Official documentation · 9 Oct 2026
Self-serve: "Simply register BytePlus to get a certain amount of free calls to the model inference service." Pay-as-you-go starts when a model is activated.
- Name
-
Official documentation · 9 Oct 2026
Product page heading "ModelArk", tagline "One-stop large language model service platform" (read in a browser). Operator: BytePlus Pte. Ltd. of Singapore (docs terms of service); its about page: "Born from ByteDance's technology".
- URL
-
Official documentation · 9 Oct 2026
BytePlus's own ModelArk product page.
- Deployment
-
Official documentation · 9 Oct 2026
Hosted API at ark.ap-southeast.bytepluses.com with a console.
- Interfaces
-
Official documentation · 9 Oct 2026
API keys from the console and SDK examples ("from arkruntime import Ark"), plus OpenAI-SDK compatibility.
- Modalities
-
Official documentation · 14 Oct 2026
Page: 'Generates high-quality images based on text and images / Generates high-definition professional videos / Quickly generate high-precision 3D assets / Accepts PDFs and replies based on document information'.
- Intended for
-
Official documentation · 9 Oct 2026
Developers and enterprises building on hosted models.
Also from Volcano Engine
-
Lumina Application
BytePlus's creative studio for image, video, audio, AI apps and an infinite canvas, running Seedream and Seedance models plus pre-built agents such as a music-video agent, sold on monthly plans to creators and enterprises producing content at volume.
-
Kickart AI agent
BytePlus's e-commerce marketing agent: give it a product link or images and it proposes concepts and a storyboard, generates the marketing video after review, produces batch ad variations from reference videos, and publishes to a connected TikTok account.
-
AgentKit Agent platform
Volcano Engine's enterprise AI-agent infrastructure platform for building, deploying and running agents on its cloud, with the VeADK and AgentKit SDKs.
-
HiAgent Agent platform
Volcano Engine's enterprise platform covering the whole agent lifecycle -- build, run, use and govern 'digital employees' -- with private deployment on the customer's own infrastructure.
-
Machine Learning Platform (机器学习平台) Platform
Volcano Engine's enterprise cloud-native machine-learning platform for training and deploying models; the platform itself is free and customers pay only for the compute resources used.
-
LLM Application Firewall (大模型应用防火墙) Infrastructure service
Volcano Engine's security service that screens the inputs and outputs of LLM applications against prompt injection, abuse, compute abuse and sensitive-data leaks, with support for custom detection models.
Open Volcano Engine in the directory
Something wrong here? Send a correction — quote this product id: byteplus-modelark.