GPT-6 Astra, Looped Transformers, and Hidden Reasoning

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1whxb4j/recurrent_looped_transformer/"> <img alt="Recurrent Looped Transformer"

<table> <tr><td> <a href="https://www.reddit.com/r/OpenAI/comments/1wyrvoc/gpt61_astra_delay/"> <img alt="GPT-6.1 Astra Delay"

arXiv:2607.10110v1 Announce Type: new Abstract: Recent work on looped language models suggests that many reasoning problems benefit from greater computational depth rather than

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1wwcne4/gpt6_astra_dots_used_blender_on_its_own_computer/"> <img alt="GPT-6 Astra Dots used Blender on its

<table> <tr><td> <a href="https://www.reddit.com/r/OpenAI/comments/1ww0lpo/built_this_game_with_astra_6/"> <img alt="Built this game with Astra 6"

<!-- SC_OFF --><div class="md"><p>After hearing rumors that GPT-6 Astra uses some form of recurrent depth, I decided to make a benchmark that specifically forces LLMs to do long

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1w4wc0p/openais_astra_uses_recurrent_depth_to_think/"> <img alt="OpenAI’s Astra uses &quot;recurrent

<table> <tr><td> <a href="https://www.reddit.com/r/OpenAI/comments/1wwq6yn/openai_publishes_a_model_guide_for_the_gpt6/"> <img alt="OpenAI publishes a model guide for the GPT-6