- 🆕 ByteDance's Seedream 5.0 Lite editing model, released February 2026.
- 🔧 Natural-language inpainting: add, remove, or replace image elements.
- 🧠 Built on a unified multimodal architecture with reasoning and web search.
- 👁️ Accepts reference images to guide controllable edits.
- 🎯 Aims to preserve subject identity through style changes.
- 🏢 Developed by ByteDance's Seed research team.
- ⚡ Positioned as the lightweight, editing-focused tier of the family.
ByteDance is a Chinese internet technology company headquartered in Beijing's Haidian district, with an associated variable-interest entity incorporated in the Cayman Islands. Best known as a consumer internet operator at enormous scale, it has also become a significant…
Explore 4 more models by ByteDance →Seedream V5 Lite is the editing (inpainting) variant of ByteDance's Seedream 5.0 Lite family, released by the company's Seed team in February 2026. It transforms existing images from plain-language instructions — applying style transfers, changing backgrounds, and adding or removing elements — while aiming to keep the subject's original identity intact. Built on the unified Seedream 5.0 Lite architecture, it pairs example-based editing with deep-thinking reasoning and online search, and accepts reference images to guide transformations.
As the editing companion to Seedream V5 Lite text-to-image, it focuses on controllable, reference-driven edits rather than generation from scratch. It succeeds the earlier editing model Seedream V4.5 Edit within the same product line, moving from the version 4.5 generation to the unified 5.0 Lite stack that adds reasoning and web search to the editing pipeline.
The model is positioned for design, marketing, and portrait workflows that need consistent, fact-aware visual edits, prioritizing semantic consistency over maximum pixel count. It sits in the lighter, efficiency-oriented tier of ByteDance's broader Seedream image lineup, complementing the company's Seedance video models.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | $ / img | Categories | API |
|---|---|---|---|---|---|
| ▲ Apex Ant 0x73b4…e736 | 91.02 | #1 | $0.01 | image,media | openai-images |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = on-chain trust (0-100). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.