Hello everyone, today I want to delve into a long-standing dilemma within the Cocos community:
Should I use Cocos Creator 3.x or 2.x for my new project?
The reason I’ve decided to write on this topic is that I’ve been receiving an increasing number of inquiries about it through private messages recently.
This might be due to the growing number of projects switching to Cocos Creator from other engines, which has led to more frequent discussions on this point.
Since the release of Cocos Creator 3.0, this question has been a source of confusion for many developers. Whether it’s on forums or in chat groups, developers often have intense discussions about this topic…
Even though Cocos Creator 3.8 has now been published as LTS (Long Term Support) version, there are still many who are struggling to make a decision.
You might often hear suggestions like:
- “Definitely go with 2.x, it’s stable!”
- “You have to use 3.x, the new features are fantastic, and you don’t want to fall behind!”
In the software field, the notion that older versions are more stable, while newer versions have more features, has become an almost reflexive judgment.
Let’s not discuss the right or wrong of these conclusions, because any conclusion is made under certain premises.
Here, Kylin can’t provide a one-size-fits-all answer, but today, combining the needs and feedback from the community developers, I’ve summarized some typical scenarios for your reference.
The following conclusions assume that everyone is working on mid-to-heavyweight products. For lightweight products, it doesn’t really matter which engine you use; you might not even need an engine at all.
Scenario 1: Mature Cocos Creator 2.x Product
Those who have worked on commercial projects know that once a company has a successful product, new projects are more inclined to be based on the source code of that product.
In addition to being able to reuse certain underlying architectures and technology stacks, there are significant advantages in terms of peripheral channel access, workflows, device compatibility, and crash rates.
These aspects require long-term accumulation and operational feedback to stabilize.
In this situation, my advice is, if Cocos Creator 2.x can meet the needs of the new project, do not switch to 3.x, nor consider upgrading to 3.x.
However, there’s another scenario I would advise against: Initiating a project with Cocos Creator 2.x for 3D development.
Although Cocos Creator 2.x has basic 3D rendering capabilities, it’s quite weak for a 3D project. To complete a 3D project, a lot of additional features need to be added. Once these 3D rendering features are added, it may not be as stable as the current Cocos Creator 3.8.x. After all, since the release of Cocos Creator 3.0.0, there have been 27 iterations of Cocos Creator.
Scenario 2: New Teams Without Source Code Accumulation
If you are a new team without source code accumulation, there’s no need to say more: it’s recommended to start directly with Cocos Creator 3.8.
Scenario 3: Teams from Other 3D Engines
If your are a team transitioning your technicial stack from other 3D engines, it’s recommended to start with Cocos Creator 3.8, whether for 2D or 3D projects, because the workflow of Cocos Creator 3.x is closer to that of mainstream 3D engines.
Scenario 4: Multi-platform Publishing
For games that require multi-platform publishing, it is recommended to use Cocos Creator 3.8, whether for 2D or 3D projects.
In Cocos Creator 3.8, the Cocos engine has perfected the dual-kernel mode, allowing products to perform at their maximum potential on native, web, and mini-game platforms.
I’m hoping the above sharing can help you guys make a better decision when you start a new project.
So, the question arises. Which version will you use in your new project?