Correction, October 13, 2026: “Confused reporting” is OpenAI chief scientist Jakub Pachocki’s phrase. OpenAI already hid most reasoning traces from users before Astra; the monitoring debate also concerns developers’ access to internal traces.
Correction, October 11, 2026: Looped transformers are Raschka’s proposed explanation for Astra’s shorter reasoning, not a confirmed description of its architecture.
A report linked GPT-6 Astra’s rumored looped architecture to concerns about hidden reasoning. Chain of thought is the intermediate reasoning text a model generates before answering. OpenAI already hid most of that text from users before Astra.
AI researcher Sebastian Raschka argues that looping does not by itself establish concealment.
What he said
Raschka quotes OpenAI chief scientist Jakub Pachocki calling the coverage “confused reporting.” Raschka thinks Astra likely uses a technique called a looped transformer, where the model runs the same set of internal layers more than once on a problem instead of building a longer chain of separate layers.
He points to a similar architecture in the open-source model Nanbeige as a working example, where “the hidden states are fed back through the same 22 blocks” a second time, doubling the effective depth of the reasoning without doubling the number of trained layers. Shorter visible reasoning, in his view, doesn’t mean the model is concealing steps. It can just mean the model needs fewer of them: “Using fewer tokens could just mean that the model is more capable and makes fewer mistakes, uses less backtracking.”
Who he is
Sebastian Raschka, PhD, is an AI researcher and author of books on building and understanding large language models, including “Build a Large Language Model From Scratch.” He writes the Ahead of AI newsletter, where he regularly breaks down new model architectures in technical detail.
What he gets right, and where it’s incomplete
Raschka’s technical distinction holds up. A model that loops through the same layers twice is doing extra computation, not hiding a reasoning trace. It is a long-studied architecture choice. He also isn’t alone in pushing back on the concealment framing: OpenAI’s chief scientist has said chain of thought monitoring remains a priority internally, and that any drop in how easily a model’s reasoning can be tracked stems from factors unrelated to this kind of architecture change.
Where the piece is thinner is on certainty. Raschka draws on the architecture report, published looped-transformer research and Pachocki’s comments. He also thinks improved training recipes and data are likely more important to Astra’s gains. OpenAI hasn’t confirmed Astra actually works this way, so this remains an informed hypothesis, not a confirmed technical fact about a closed model.
Why it’s notable
Astra’s capabilities and Copilot pricing are one question. Whether builders should trust what it shows them while it works is another. That question matters more as coding agents make more autonomous decisions during long tasks, and users increasingly rely on visible reasoning to catch a mistake before it ships.
What it means for builders
Don’t treat a shorter reasoning trace as evidence a model is hiding something from you. Trace length alone cannot distinguish efficient problem-solving from less informative reasoning. Raschka offers possible mechanisms without establishing which architecture Astra uses.
The practical takeaway is to keep judging a model like Astra by its output quality and by whether it still lets you inspect its work when something goes wrong, not by the raw length of the reasoning it shows you.
End of article