Midjourney V8

What V8 is

Midjourney V8 is a version of Midjourney's text-to-image model, first opened to users as an alpha on 17 March 2026 at a separate address, alpha.midjourney.com^[1]. As with earlier releases, a user writes a prompt and the model returns images. The alpha label means it was published for use before being treated as a finished release.

Speed and new render modes

The headline change at launch was speed: Midjourney said the new version generates images roughly five times faster than before^[1]. Two new options change how an image is rendered. A new --hd mode renders natively at 2K, and a --q 4 mode adds coherence^[1]. Both were costly at launch: jobs using them cost four times as much as ordinary jobs^[1].

Text inside images

Midjourney said text rendering "works better than ever" when the words to be drawn are placed in quotation marks^[1]. In prompting practice, quotation marks are a common way to mark words that should appear in the picture rather than be read as part of the scene description.

Compatibility with existing controls

V8 supported multiple aspect ratios and kept the prompt parameters --chaos, --weird, --exp and --raw^[3]. It also remained backwards compatible with V7 personalization profiles, moodboards and style references^[3], so the tuning users had already done to steer the model toward their own taste carried over into the new version. Those style reference and moodboard jobs came at a price at launch: they were four times as slow as regular jobs, and each cost four times as much^[3].

Changes to the web interface

On the web side, the alpha brought an improved conversation mode, a Grid Mode for viewing one large set of images at once, and settings moved into sidebars^[3].

Limits at launch

Relax mode was not supported at launch^[1].

Follow-up versions

Midjourney followed the alpha with V8.1, itself an alpha, on 14 April 2026, and V8.2 on 24 July 2026^[2]. The V8 line arrived in the same year as Niji 7, the anime-focused model released on 9 January 2026^[2], so the company updated its general-purpose and anime-oriented models within a few months of each other.

Where it sits

Taken together, the changes in V8 are about throughput and control rather than a new kind of output: faster generation, a native 2K option, a mode for coherence, better text rendering, and continuity with existing parameters and personalization settings. For people who use the model inside a daily workflow, those details decide how many variations can be tried in a given stretch of time, and whether style settings built up on the previous version still apply. The quick succession of V8.1 and V8.2 shows the version being revised in public rather than frozen at the alpha.