Playco正在使用GPT-6 Astra构建Playbot，这是一款面向专业游戏开发者的AI驱动IDE。它直接连接Unity和Godot等引擎，使AI模型能够编辑场景、运行和测试游戏、验证更改，并在开发者已使用的工具中并行工作。
在游戏开发中，模型需要做的不仅仅是编写代码。它必须推理空间、视觉参考、响应式界面、游戏手感，以及更改在游戏实际运行时是否有效。Playbot通过让模型在引擎中构建、运行测试、发现错误并改进其创建的内容，使这一切成为可能。
“GPT-6 Astra在空间推理和元素定位方面表现出色，结果非常合理。此外，其视觉能力似乎也有所提升。我们还看到了游戏引擎中UI响应性的改进。”
——Joao Vieira，Playco首席产品工程师
从一个灰盒构建三个可玩的游戏世界
使用GPT-6 Astra，Playco团队从简单图元创建了一个无主题的灰盒原型，然后在游戏玩法和创意细节上进行了几次迭代。基于这一共同基础，团队开发了三个主题游戏原型。Playco表示，GPT-6 Astra一次性生成了全部三个原型。
大多数原型在首次尝试时即可运行。“使用Astra，第一个原型已经非常出色。我们唯一需要做的更改是基于我们的游戏偏好，”首席产品工程师Joao Vieira表示。其中一个赛博朋克版本需要性能修复，但其余部分无需额外迭代即可完成。
手动修复减少一半
使用之前的模型时，Playco的初始灰盒不够精细。提示模型进行修正最终适得其反，因此工程师不得不手动介入修复游戏。使用GPT-6 Astra后，Playco报告手动修复减少了50%。
团队还观察到空间推理、参考图像重现、Unity内响应式UI以及游戏手感方面的改进。由于Playbot可以让模型玩游戏并验证自身更改，Playco表示Astra还能更轻松地发现错误，并识别出改善玩家体验的环节。
对Playco而言，最大的影响是能够将更多创意转化为开发者可以实际游玩和比较的内容。正如首席产品工程师Joao Vieira所说：“如果你有10个游戏创意，你可以全部实现，并实际游玩它们，感受它们的体验，而不仅仅是凭空想象。”
Playco is using GPT‑6 Astra in building Playbot, an AI-powered IDE for professional game developers. It connects directly to engines such as Unity and Godot so AI models can edit scenes, play and test games, validate changes, and work in parallel inside the tools developers already use.
In game development, a model needs to do more than write code. It must reason about space, visual references, responsive interfaces, game feel, and whether a change actually works when the game is played. Playbot makes that possible by letting the model build in the engine, run tests, find bugs, and improve what it created.
“GPT-6 Astra is much better at reasoning about space and positioning elements in a way that makes sense. Also, its vision capabilities seem to be improved. We also saw improvements in UI responsiveness in game engines.”
—Joao Vieira, Lead Product Engineer, Playco
Building three playable worlds from one grey box
Using GPT‑6 Astra, the Playco team created an unthemed grey box prototype from simple primitives, and then made a few iterations on gameplay and creative details. From that shared foundation, the team developed three themed game prototypes. Playco says GPT‑6 Astra produced all three in one go.
Most prototypes also worked on the first take. “With Astra, the first prototype was already strong. The only changes we needed to make were based on our gameplay preferences,” said Joao Vieira, Lead Product Engineer. One cyberpunk version needed a performance fix, but the rest came together without additional iteration.
Cutting manual fixes in half
With the previous model, Playco’s initial grey box was less polished. Prompting the model to correct it eventually became counterproductive, so engineers had to step in and fix the game by hand. With GPT‑6 Astra, Playco reports 50% fewer manual fixes.
The team also saw improvements in spatial reasoning, recreating reference images, responsive UI inside Unity, and game feel. Because Playbot can let a model play the game and validate its own changes, Playco says Astra also found bugs more easily and identified places to improve the player experience.
For Playco, the biggest impact is being able to turn more ideas into something developers can play and compare. As lead product engineer Joao Vieira puts it, “If you have 10 ideas for a game, you can do all 10 and actually play them and see how they would feel rather than just imagine.”
本文内容采集自官方网站，排版和翻译可能与原页面存在差异。
阅读官方全文