OpenAI's latest model release with prompt-length and token-spend debate.
Source discussion: GPT-5.6
Editorial paraphrase
Replies compared GPT-5.6 with Codex and Claude Code, and debated whether shorter prompts really save tokens.
Read the original thread