MiracleVision 美图奇想
Meitu's aesthetics-led vision model, now image, video and design
Overview
MiracleVision is Meitu's in-house AI vision model, powering image generation, AI design (vector art, text effects, layout) and video across Meitu's own apps and an enterprise API. Aimed at e-commerce, ad and film teams that want production-ready visuals fast.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Text-to-image
- Image-to-image editing
- Same core API
- Volume discount
- Bulk pricing
- SDK/SaaS/custom model training options
What you can produce with MiracleVision 美图奇想
- Text-to-image and image-to-image generation
- AI design tools: vector graphics, text effects, smart layout
- AI video generation (MiracleVision V6)
- Ranked #1 on VBench video-generation benchmark (85.30%)
- Enterprise API/SDK/SaaS access
- Custom model training for partners
- Powers Meitu consumer apps (美图秀秀, Wink, MOKI)
ASEAN Perspective
MiracleVision 美图奇想 in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
MiracleVision's strongest argument is provenance: Meitu has spent over a decade tuning image processing for a billion-plus user base, and that aesthetic instinct shows in cleaner composition, more natural lighting and better skin and fabric texture than most general-purpose generators. The V6 model launched in June 2026 folds text, image, video and audio into one routing system, and it topped the VBench video-generation leaderboard with an 85.30% overall score — a third-party-verified result, not just marketing copy.
The catch is distribution. Access outside China runs through the API/SDK route rather than a polished international product, English documentation is thin, and independent review coverage (G2, Capterra, Trustpilot) is essentially nonexistent — most of what's written about it lives on Zhihu and Chinese tech press. Treat it as a capable backend for e-commerce, ad and film production pipelines that can absorb a China-facing integration, not a drop-in consumer tool.
What people say
MiracleVision (internally 奇想智能, Qixiang Intelligence) is Meitu's self-developed visual model, first opened to the public in early 2024 and rebuilt as a multimodal mixture-of-experts system — MiracleVision V6 — unveiled at Meitu's Multimedia Festival in June 2026. The pitch shifted from a text-to-image tool into what Meitu calls "delivery-based AI": eight linked products spanning image generation, AI design (vector graphics, text effects, smart layout) and video, all routed through one model.
On the benchmark side, the results are independently checkable: MiracleVision's video model placed first on VBench, a widely-used third-party video-generation leaderboard, scoring 85.30% overall and 86.69% on the video-quality sub-metric — ahead of several better-known international models at time of testing.
User-side commentary is thinner and mostly in Chinese. One widely-shared Zhihu writeup from a commercial portrait photographer described the tool as functioning like a "new colleague" in a professional retouching workflow — useful for batch consistency and quick concept iterations rather than one-off hero shots. Chinese tech outlets (AIbase, GeekPark and others) have tracked the model's version history closely, from V1 in 2023 through V6 in 2026, and consistently frame it as Meitu's core differentiator against Adobe Firefly and Midjourney in the domestic market.
What's missing is independent Western review coverage — no G2, Capterra or Trustpilot listing turned up in search, which makes it hard to benchmark real-world reliability outside China the way you can for its competitors. Commercial access runs through Meitu's open API: published rates are ¥0.1 per image under 100,000 images a month, dropping to ¥0.08 between 100,000 and 1,000,000, and ¥0.06 above a million, alongside SDK, SaaS and custom model-training options for enterprise partners. For teams already inside Meitu's consumer apps — 美图秀秀, Wink, MOKI — MiracleVision is the generation engine running underneath rather than something they'd install separately.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including MiracleVision 美图奇想's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with MiracleVision 美图奇想 unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to MiracleVision 美图奇想 directly →
Spotted something out of date? Suggest an update →
MiracleVision 美图奇想 in the news
AI & ML
ByteDance Launches Seedream 5.0 Pro, an Image Model That Outputs Editable Layers
Cybersecurity
Two Rival Powers, One Police Force: Researchers Say China- and India-Linked Groups Separat...
Cybersecurity
Taiwan Charges Two Executives Over a LINE-Account Leasing Scheme Prosecutors Link to Chine...
Alternatives to MiracleVision 美图奇想
More in Image Generation
