Midjourney rolls out V6.1 with sharper details and 25% faster speeds
Midjourney just launched V6.1 as its new default image model. The update delivers cleaner hands, faces, and textures, smarter text rendering, and a 25% boost in standard generation speeds, alongside an upgraded 2x upscaler. Inpainting and zooming still temporarily fall back to V6.0.

Midjourney just launched V6.1, rolling out a major mid-cycle upgrade to its flagship image generation model with a laser focus on fixing AI art’s most persistent headaches.
The new model is live right now and has immediately replaced V6.0 as the default engine across the entire Midjourney platform.
Smarter limbs and cleaner pixels
The biggest upgrade in V6.1 comes down to visual coherence. The model does a much better job generating correct human and animal anatomy, including arms, legs, hands, and bodies.
Small, distant features that used to turn into mushy noise are also significantly sharper. Tiny faces, distant eyes, and far-away fingers now retain their proper structure.
Everyday image quality gets a noticeable bump as well, with reduced pixel artifacts and cleaner rendering of skin and complex textures.
Generation speeds for standard image jobs are roughly 25% faster. Midjourney also packaged in a new 2x upscaler that delivers cleaner texture quality than previous options.
Text rendering sees another boost. Putting words inside quotation marks in your prompts will now result in far more accurate in-image typography.
Personalization codes and a few caveats
The update introduces an overhauled personalization engine designed to offer better nuance, surprise, and stylistic accuracy.
Power users also get personalization code versioning. That means you can pull any personalization code from an old generation to faithfully replicate that exact style profile in a new job.
Midjourney also added a new --q 2 command. It adds about 25% to render times to push for deeper surface texture, though the team warned it can sometimes degrade overall coherence.
There is one key blind spot: V6.1 currently lacks its own inpainting and outpainting models. If you use zoom, reframe, or vary region, Midjourney silently falls back to V6.0 under the hood.
The team noted this update was guided directly by community votes on its ideas board, and a V6.2 release could drop within the next month once enough user data rolls in.
Why it matters
If you use AI imagery for client work, concept art, or game design, V6.1 is an immediate no-brainer upgrade.
A 25% speed bump paired with fewer deformed hands and melted distant faces saves real hours in post-production and re-rolling.
We think Midjourney’s incremental release strategy is smart. Instead of making users wait half a year for a massive V7, shipping practical community-requested fixes keeps paying subscribers engaged.
However, the lack of native V6.1 inpainting is a legitimate frustration. Having an edit step fall back to an older, less capable model introduces a jarring visual mismatch right when you need precision most.
Midjourney remains the aesthetic gold standard for generative art, but as open-source competitors close the gap, seamless editing tools are going to matter just as much as raw image beauty.
Sources (2)
- twitter.com(opens in a new tab)Originaltwitter.com
- updates.midjourney.com(opens in a new tab)Reportupdates.midjourney.com