Seedream 4.0
ByteDance's 4K unified image generator and editor — topped the Artificial Analysis image arena at launch.
About
Seedream 4.0 is ByteDance Seed's multimodal image model, launched September 9, 2025. It unifies text-to-image generation and image editing in one system: 4K output (2K in ~1.8 seconds), up to six reference images for precise edits, visual signal control, and in-context reasoning. It is live on Dreamina, the Doubao app, and Jimeng AI, with enterprise access through Volcano Engine and third-party endpoints on fal.ai and Replicate (roughly $30 per 1,000 images).
Independently, Artificial Analysis ranked Seedream 4.0 #1 on both its Text-to-Image and Image Editing arenas at launch with an Elo of 1205 — just ahead of Google's Gemini 2.5 Flash (Nano Banana) at 1201. ByteDance's 10x-speed-over-predecessor claim is provider-reported. Like all top image models, its photorealism raises deepfake and misinformation concerns; outputs should be treated accordingly.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Independent (Artificial Analysis, Sept 2025): #1 Text-to-Image and Image Editing arenas at launch, Elo 1205 vs Gemini 2.5 Flash (Nano Banana) 1201.
Provider-reported: 10x faster than Seedream 3.0.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Release and launch specs confirmed across independent press (TechRadar, KrASIA) plus an independent #1 arena ranking from Artificial Analysis. Docked for: (1) proprietary, API-only — no weights; (2) speed and quality superlatives partly rest on provider-reported figures; (3) photorealism makes it a capable deepfake instrument, a real misuse consideration.