Seedream 3.0
ByteDance's bilingual text-to-image model with native 2K output, launched April 2025.
About
Seedream 3.0 is ByteDance Seed's text-to-image foundation model, announced in April 2025 with a technical report dated April 15. It is a Chinese-English bilingual generator focused on prompt alignment, fine-grained typography (including complex Chinese characters), and native high-resolution output up to 2K.
The team reports a 4 to 8 times speedup over Seedream 2.0 via a new acceleration paradigm while generating a 1K image in about three seconds under its test conditions. It is integrated into ByteDance's Doubao chatbot and the Jimeng creative platform, and was available through Volcano Engine as Doubao-Seedream-3.0-t2i. A unified successor, Seedream 4.0, followed later in 2025.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 114; Trending score 82/100 from researcher assessment.