UE5 needs to get finished. Why we need UE 5.9

[OCEAN_STORM]

Saying that UE 5.8 is the last planned major Unreal Engine 5 release feels like leaving UE5 unfinished. Many developers are not asking for a bigger engine, a new generation, or another wave of experimental features. We are asking for the current UE5 generation to be finished and optimized.

The main problems are very clear:

• Many important UE5 features are still Experimental, Beta, or not realistically production-ready for many projects.
• UE 5.8 also introduces major new systems such as Mesh Terrain, but Mesh Terrain is Experimental. If UE 5.8 is the final major UE5 release, then developers may be left with important UE5-era features that exist in the engine, but do not have a clear path to becoming production-ready inside UE5.
• UE5 still has real performance concerns across many types of projects. UE 5.8 includes important performance work, and that is appreciated. But developers need predictable real-world performance across actual migrated projects, large worlds, production content, and normal development workflows — not only selected benchmark scenarios.

For many teams, especially game developers, UE6 is not the practical answer right now. We need a stronger UE5.

A focused UE 5.9 would be extremely valuable if it were positioned not as a feature-expansion release, but as a stabilization and production-readiness release:

• Improve performance across real projects.
• Fix major regressions.
• Move mature Experimental/Beta features to Production-Ready status where possible.
• Give developers a final, reliable UE5 version that can be trusted for shipping games over the next several years.

The concern is the gap. If UE 5.8 is the last planned major UE5 release, and UE6 Early Access is still targeted for the end of 2027, with the full UE6 release coming later, developers may be left for years in an awkward position: UE5 is the engine we are building and shipping with today, but many powerful UE5 systems may remain unfinished, experimental, unclear, or risky for production.

If UE 5.8 were already fully stable, highly optimized, and most core features were production-ready, this would be much easier to accept. But that is not where many developers feel UE5 is today.

Please reconsider and commit to at least one more major UE5 release — UE 5.9 — focused entirely on performance, stability, regressions, production-readiness, and maturing the systems already introduced in UE5.

UE5 does not need to get bigger. It needs to get finished.

[ClockworkOcean]

I agree wholeheartedly… also dropping blueprint is a bad idea, see

They want to remove Blueprint in UE6

@Yaeko Fancy making this vote-able? I second, third, fourth etc this sentiment. Getting rid of BP is a catastrophically crap idea… they spend all this time modding the engine ( without really working on efficiency ) to on-board loads of creators, only to cut them off at the last second. Nobody wants to develop for film/TV in a fortnite engine… Creatives don’t want to code. ( that’s a big full stop there )

[ChrisK]

We can always have 5.8.x. It doesn’t mean that they will leave it unfinished. I take it that there will be no new major features but FINISH it as it is meant to be.

IMHO, UE as is will not survive in AI era and I felt their urgency to make it more AI friendly as soon as possible.

First removing, AI unfriendly BP and replace it with Verse.
Yes, there will be Visual Verse so BP is not going away but coming back face-lifted.

So, I think I get a very positive vibe overall. Wish that they made little clear about it.

[BDC_Patrick]

Lumen - Crap of performance. haven’t tested Lumen Lite, yet. But If it is the same bad experience in setting up/managing the system as with normal lumen - I consider it absolutely unstable and not in the nearest finished. Deactivating lumen alone is a pain, since there is no UI to handle it!

Nanite - hahahahaha… Let’s not start with that. So much stuff is missing from normal meshes to nanite meshes. Starting with full Translucency support and the same Displacement / Tesselation performance than default Meshes. Nanite on Mass with Lumen kills performance.

MassEntity - Talking about Mass. Promising System. I was so ■■■■ mad at the Epic staff that created it, since it killed my nerves for over 5 months trying to set it up… because it is the absolute worst system to deal with in Unreal. No UI, the framework is minimum documented, tutorials only cover the bare minimum (mostly “how to add 20 GTA NPCs”). Nothing is clear and spot-on plausible when trying to set this up.

Megalights - Production ready.. sure.. if you say so.. would say “Beta”. Performance still is A$$.

CommonUI - As someone dealing with UI / UX every Day I can say (heartly), CommonUI is in Beta Stage.. nothing more.. So many functionality not exposed to BP, so much not documented or even halfway completed. The promised 360° focus handling has been removed from the roadmap with 5.7 - simply because epic knows they will not finish this up..
A SHAME!

Multicore Cooking - If you can make it work, good for you. We tried here.. lots of time. It never worked and always threw Error after Error. Maybe sounds good on paper, but is badly implemented!

PGC - I tried for some months, and still cannot get anything good from it. it simply is not documented and integrated well enough. With Mesh terrain, even less.

Mesh terrain - It is new.. VERY NEW.. so.. maybe 50% completed in 5 Years or so.. cause Epic games.. maybe never completed.. like Landscapes..

Substrate - Is still super complicated to handle and lets’s performance drop quickly if done wrong. Much more documentation and tutorials of REALLIFE situations is needed..

Documentation - It’s Epic games.. I gave up on this 10 years ago..

Virtual Shadow Maps - Say goodbye to 60+FPS

Metahumans in game - See VSM

Enhanced Input support in UI - Workaround to make it work with duct tape.. not really integrated and unintuitive..

Just to get the core features of UE5

And I bet: They won’t get all of them fixed.

UE6 will start its life with the remaining Legacy issues and fixing-corpses of UE5..

And on top of that, they want to remove BP in later UE6.. nice.. more ■■■■ that won’t get fixed or work in a realistic performant way, within the next Decade..

Epic should stop throwing 10 new Features with every release onto the Devs, and start actually working on fixing Code which is buggy for years now!!

Give the Devs some intuitive, performant and WORKING Engine - Instead of Frankenstein’s monster of workarounds, glued together with the best faith and pleasing words!

PS.:
I really hope that within the next years an actual Fortnite killer will spawn on steam or such, to get rid of this stupid game.. so epic can finally focus back on making the engine good, instead of pleasing Shareholders with more Fortnite integrations no one wants or has even asked for.

[pyrsen1]

wdym there is no ui to handle it you can easily disable lumen i the project setttings

[IronicParadox]

@BDC_Patrick

Deactivating lumen alone is a pain, since there is no UI to handle it!

Easily done with either a post process volume or within the project settings.

@BDC_Patrick

Lumen - Crap of performance

It runs pretty well and is definitely in decent shape these days. Still has room for improvement on denoising though.

@BDC_Patrick

Nanite - hahahahaha

Works very well. You shouldn’t really be needing translucency support for it though. If you have a glass window, it should be a regular mesh. If you have leaves on a tree, they should be actual geometry or at the very most, masked. The new voxel system works great as well, though it could use some kind of means of translating WPO effects to the voxels (currently, it just causes a weird distortion).

@BDC_Patrick

MassEntity

Experimental, so a lot of it is not fully implemented yet or documented.

@BDC_Patrick

I wanted to kill the Epic staff that created it, since it killed my nerves for over 5 months trying to set it up

Extremely childish and not the kind of language you should ever use in an online environment. It’s not even something that should be joked about.

@BDC_Patrick

Megalights

It’s also in a pretty good place these days as well and the performance is fine. It’s meant to be used when you have a lot of lights in a scene.

@BDC_Patrick

PGC

PCG works extremely well and is definitely production ready. The fact that you “still cannot get anything good from it” isn’t a problem with the system, it’s a problem on your end. Get on the “Unreal Source” Discord server and check out the PCG section. There are tons of people working on awesome things there, as well as a lot of people that help others out.

@BDC_Patrick

Mesh terrain

Personally, I’ve just started messing around with it, so I’m not 100% sure how buggy or functional it is yet. Early testing impressions: It seems to work well enough and can work with PCG.

@BDC_Patrick

Substrate - Is still super complicated to handle and lets’s performance drop quickly if done wrong.

The terms and variables are an industry standard, but it does take a little getting used to the transition. It’s extremely powerful and has performance parity with the legacy material system, assuming a single shading model used. Even when using non-vertically layered materials with different shading models, the performance is still great, considering it’s impossible to do with the previous material system without writing your own shading model.

@BDC_Patrick

Virtual Shadow Maps

Meant to be used hand in hand with nanite. And FPS count is a useless metric. You say 60+fps, from what? Going from 300fps to 240 isn’t that terrible of a jump, compared to going from 90 to 30. The engine targets 60 and 30fps.

Overall, a lot of your complaints sound more like an issue with your experience or knowledge in the various systems, how they work and how they affect the performance ceiling/floor of a project. UE5.8 will likely get a few smaller patches before UE6, so some of those systems might finalize or get a little better.