GPT-6 Astra, Looped Transformers, and Hidden Reasoningkept by varun • sep 09GPT-6 Astra achieves top coding and multimodal scores via looped transformers that reuse weights for extra depth, while recent work shows this architecture does not inherently conceal chain‑of‑thought reasoning.AboutOpenAI · Mixture-of-Recursions · Universal Transformer · GPT-6 Astra · Nanbeige4.2-3B · OuroFiled#ai#ai-research#developer-tools#generative-ai#machine-learningRelatedGPT-6 Astra, Looped Transformers, and Hidden ReasoningThe article examines GPT‑6 Astra’s performance, explains looped transformers as weight‑shared depth increase, and evaluates claims that they hide chain‑of‑thought reasoning.also on Mixture-of-Recursions, Ouro, Universal Transformer, GPT-6 Astra, OpenAI, #ai, #developer-toolsWe've never seen this before. The biggest jump in Vending-Bench history. GPT-6 Astra is better at making money and more ethical than ClaudeGPT-6 Astra outperforms Claude Fable 5.1 in both profit generation and ethics on the Vending-Bench benchmark, marking OpenAI's first #1 ranking.also on GPT-6 Astra, #machine-learning, OpenAI, #ai, #developer-toolsOpenAI’s AGI number came from a harness, not the modelOpenAI’s 99.9% AGI score for GPT‑6 Astra relied on its custom Provider Adapter harness, while the benchmark’s standard harness yields only 62.7% for the same model.also on GPT-6 Astra, #ai-research, OpenAI, #ai, #developer-tools
GPT-6 Astra, Looped Transformers, and Hidden Reasoning