ByteDanceByteDance·🖌️ Inpainting

Seedream V4.5

anonymized
Try on Venice.ai ↗
Quick reference
Seedream V4.5 — TLDR
  • - 🏢 ByteDance's image-editing variant in the Seedream 4.5 family.
  • - 🔧 Instruction-driven editing of existing images via natural-language prompts.
  • - 🎯 Improved portrait quality and subject-identity retention across edits.
  • - 📚 Improved typography and dense-text rendering over Seedream 4.0.
  • - ⚡ ByteDance reports all-round gains from scaling on internal MagicBench.
  • - 📅 Listed with a 2026 release on this catalog.
💰 Pricing
$0.050
per edit
📅 On Venice since
Jan 4, 2026
150 days ago
Provider

ByteDance is a Chinese internet technology company headquartered in Beijing, widely known as the parent company behind TikTok and Douyin. Beyond social media, ByteDance has invested heavily in artificial intelligence research, building generative media models…

Read full profile →
12 models on Venice
8 video · 2 image · 2 inpaint
Since Nov 5, 2025

About this model

Seedream V4.5 Edit is the image-editing variant of ByteDance's Seedream 4.5 model, focused on precise, instruction-driven editing of existing images guided by natural-language prompts rather than full regeneration. It works as the editing counterpart to the text-to-image companion Seedream V4.5, letting users modify assets while keeping a project's visual direction.

Compared with its same-family predecessor Seedream 4.0, ByteDance reports an all-round improvement from scaling the model — better prompt adherence, alignment and aesthetics on its internal MagicBench evaluation. The vendor also highlights sharper typography and dense-text handling, plus improved portrait quality and subject-identity retention across edits. These gains target e-commerce, branding, editorial and key-art workflows where edited assets must stay faithful to source images.

Within ByteDance's broader lineup, Seedream V4.5 Edit sits alongside the lighter Seedream V5 Lite editing variant and its Seedream V5 Lite generator, while the Seedance series covers video generation. As with all releases here, capabilities depend on the deploying provider's configuration and licensing.

This About section is AI-generated from public sources (Claude Opus 4.8), with no human editing. It may contain inaccuracies — verify critical details against the sources listed above.

Data sources: Venice API · HuggingFace · Wikipedia — enrichment updated 1d ago