Reve is a text-to-image and editing model by Reve AI, whose current flagship Reve 2.1 (July 2026) generates native 4K images through a layout-first architecture — planning each image as a structured, editable layout before rendering — with best-in-class typography and precise, region-by-region control.
Built for design and marketing teams, it turns briefs into posters, packaging, and campaign visuals — with an "images as code" workflow where you adjust one element without re-rolling the whole picture, and in-image text that comes out correctly spelled.
Posters, packaging, campaign layouts, typography-heavy design, brand graphics
Design and marketing teams that need exact composition and correctly-spelled in-image text
Images as editable code — layout-first control and typography that ships correctly spelled.
Reve inserts a structured, code-based layout between your intent and the final pixels. Every element carries a position, size, and description you can edit independently — so instead of re-rolling until the composition is right, you adjust it directly. For teams building real creative products, that's a fundamentally different workflow.
Reve runs a dedicated typography pass, and it shows: headlines, packaging labels, and small captions come out crisp and correctly spelled far more often than in competing models. For posters, packaging, and any design where readable text matters, this is the feature designers notice first.
Reve generates true native 4096×4096 — 16 megapixels in a square frame — rather than a low-resolution render followed by an upscale pass. That distinction matters most exactly where upscalers fail: text, fine textures, and repeated patterns stay clean.
Reve 2.1 adds visual reasoning on top of the layout engine — greater prompt understanding and world knowledge — so it interprets a campaign brief or a pattern system rather than approximating it. It's among the best models at following complex prompts literally.
Reve 2.1 specifically improved foreign-text rendering — a rare enough upgrade that it's worth testing on real, non-English copy rather than English-only prompts — making it practical for localized packaging and campaign work.
Reve AI fits wherever teams turn briefs into precisely composed, text-accurate visuals.
Add a node, describe your idea, and generate native-4K visuals with editable layouts.
Add a Reve node to the canvas and describe your idea in a text prompt — the more specific the composition and text, the more Reve's layout planning pays off. Optionally upload an input image for style remix, and set aspect ratio and resolution.
Run the prompt. Reve plans the layout, then renders at native 4K. Put exact on-image text in quotes so the typography pass has a script, and describe the placement of each element for the layout engine.
Refine by adjusting individual regions rather than re-rolling, chain the output into an upscaler or background-removal node, or pass a clean frame to an image-to-video model, then export with commercial rights on your plan.
Real outputs with the exact prompts — copy any prompt and run it yourself.
One subscription across 30+ AI models — no per-tool credit balances or separate signups. Credit cost per generation is shown live in the node before you run it.
500 weekly credits (about 9 image generations) to test the model and see the output quality. No credit card required.
45,000 monthly credits with cheaper per-credit pricing, commercial rights, and image download.
90,000 monthly credits, up to 10 seats, shared workspace, centralized billing, and priority support.
210,000 monthly credits, API and integration support, unlimited seats, and dedicated management.
Full specs for the Reve model family.
How Reve compares with other image models in the Phygital+ catalog.
Chain generation, editing, upscaling, and video into one repeatable workflow.
Generate with Reve, then upscale to print-ready delivery resolution in one workflow.
Generate product visuals and remove or replace backgrounds for e-commerce listings.
Pass a clean Reve reference frame into a video node to animate it for social and ads.
Have a text model write the detailed layout brief, then generate the visual with Reve.
Everything you need to know about Reve in Phygital+.
Reve is an AI image generation and editing product from Reve AI, a Palo Alto lab. Its current flagship, Reve 2.1, released July 9, 2026, generates native 4K images through a layout-first architecture — planning each image as a structured, editable "images as code" layout before rendering — with a dedicated typography pass for crisp in-image text.
Usage is credit-based and included in every Phygital+ plan — from the free plan up to Enterprise. The exact credit cost per generation is shown live in the node before you run it, so you can start for free and scale as you go. This is separate from Reve's own pricing, which offers a free tier plus paid plans at roughly $7.99 and $19.99/month.
Outputs created on any paid plan (Starter and up) come with commercial-use rights. Note that Reve applies a fairly locked-down content policy on its own platform. Check the model card for any provider-specific limits.
Reve is made by Reve AI, a Palo Alto startup founded by former Google Brain and NVIDIA researchers. Its first model, Reve Image 1.0, reached the top of the Artificial Analysis Image Arena; the team is small — roughly 65 people — competing against far larger labs.
Layout-first generation inserts a structured, code-based layout between your prompt and the final image. Every element carries a position, size, and description you can edit independently, so an image is addressable like code — you adjust one region instead of re-rolling the whole render.
Reve generates native 4K — 4096×4096, about 16 megapixels — rather than a low-resolution render followed by an upscale pass. Native generation keeps text, fine textures, and repeated patterns clean, which is exactly where upscalers tend to fail.
Yes — it's one of its strongest features. Reve runs a dedicated typography pass, producing crisp, correctly-spelled headlines, packaging labels, and captions more reliably than most competitors. Reve 2.1 also improved foreign-text rendering for localized work.
Reve 2.1, released July 9, 2026 — a month after Reve 2.0 — improves prompt understanding, world knowledge, and foreign-text rendering, while continuing the focus on native 4K generation, editable layouts, and precise typography. On the Text-to-Image Arena it ranks #2 at an Elo of 1306.
Yes. Reve offers per-credit API access to its hosted models (it's proprietary, not open-weights). In Phygital+, you can run Reve as a node on any plan, and API and integration support is available on the Enterprise plan.
Running Reve in Phygital+ means no separate signup or credit balance for one model: it sits on the same canvas as 30+ other AI tools. Chain generation, editing, upscaling, and video into one repeatable workflow under a single subscription.
Browse the full catalog of 30+ AI models available in Phygital+.
Join 100+ teams using Phygital+ – every model in one workspace
Try Phygital+ free
