Seed 2.0 Pro
Language modelCard summarizedFlagship general-purpose agentic model
Source: Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity (87 pp)
Use it via API
Where to call it
- BytePlus ModelArk
seed-2-0-pro-260328
Supports
- Streaming
- Reasoning control
- Tool calling
- Batch API
- Image input
- Prompt caching
Good for
- Search and deep-research agents (BrowseComp 77.3, HLE-text with tools 54.2)
- Competition maths and programming (AIME 2025 98.3, Codeforces 3020)
- Visual reasoning on charts, documents and screenshots
Overview
Seed 2.0 Pro
Flagship general-purpose agentic model
- Cost
- $0.50 in
- Context
- 256K
- Input / output
- $0.50 / $3.00
- Max output
- 128K
- Knows up to
- —
- Released
- 14 Feb 2026
Seed2.0 is ByteDance Seed's production model family (Pro / Lite / Mini) tuned for large-scale deployment: visual and multimodal understanding, fast inference options, and reliable execution of complex instructions. In ByteDance's own tables Seed2.0 Pro is level with GPT-5.2 High and Gemini 3 Pro on maths, competitive programming and search agents, but the card states it still has 'considerable gaps with Claude' on coding and 'relatively obvious gaps with Gemini' on long-tail knowledge. §1, p.2
seed-2-0-pro-260328- Input / output per 1M
- $0.50 / $3.00
- Context window
- 256K
- Max output
- 128K
- Knows up to
- —
- Reasoning
- Deep reasoning on by default; can be disabled; reasoning_effort minimal → max
- Weights
- Closed
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses ByteDance Seed's Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity. Every claim carries its source page.