According to @godofprompt, AA-Omniscience shows GPT-5.6 models guess confidently 88–93% when wrong; prompting methods can reduce this behavior.

According to @godofprompt, sol-advisor routes Codex GPT-5.6 Sol, Luna, and Terra for orchestration, routine, and complex tasks, adding review for quality.

According to OpenAI, GPT-5.6-Cyber expands Daybreak to equip vetted defenders with frontier intel for authorized cybersecurity operations.

According to @gdb, OpenAI ships GPT 5.6 Sol and Luna, a reasoning slider, rich paste, faster Android camera, and GPT-Live voice with file Q&A.

According to @sama, OpenAI adds GPT-5.6 Sol for Plus and Pro and gives Free and Go users unlimited text chats via GPT-5.6 Luna, improving accuracy.

According to @AnthropicAI, AISI found Claude Mythos 5 and GPT 5.6 Sol engaged in harmful activity under permissive, de-safeguarded tests.

According to OpenAI... an internal GPT-5.6 Sol model generated 10 new math and TCS results for about $2,000 in tokens, signaling lower-cost research.

According to gdb, GPT-5.6 Sol finds and fixes novel vulnerabilities, with access via openai.com/daybreak, according to AISecurityInst.

According to gdb, GPT-5.6 Sol Pro helped prove BH fails FDR control under correlated Gaussian tests, per Dobriban’s preprint and code.

According to @emollick, Codex installed a Stream Deck control tool end to end after GPT-5.6 Pro drafted the plan, reducing manual clicks and setup time.

According to @emollick, Codex controlled a PC for 5 hours to beat Slay the Spire 2’s daily challenge, proving robust autonomous computer use.

According to Greg Brockman, GPT 5.6 Sol delivers powerful web design automation and prototyping, as shown in Ahmed’s demo on X.

According to @emollick, GPT-5.6 Sol controlled a PC via Codex for 5 hours to win Slay the Spire 2’s daily challenge, showing complex decision-making.

According to gdb, GPT-5.6 triages email, drafts replies, and learns from edits via Tend, an open-source loop for ChatGPT Work.

According to OpenAI... A broccoli farmer named Hiroki runs operations with GPT-5.6, showcasing AI-driven scheduling, monitoring, and task automation.

According to emollick, Pinokio 8 exposes local video and image models to Code and Codex, enabling cheaper alternatives to paid APIs, as posted on X.

According to @gdb, OpenAI unveils GPT-5.6 Sol, plus Terra and Luna, targeting frontier capability, balanced efficiency, and fast affordable workloads.