@OpenAI Welcome! But GPT-6 Sol burns usage MUCH faster than GPT-5.6 Sol. 14% in a few hours! (Even when using Luna for subagents. Pro $100 plan.) I thought the new Sol and Luna were supposed to be cheaper...
@thsottiaux Reasons I moved:
- Codex got code right first time (gpt 5.5, less so gpt 5.6)
- $100 subscription
Reasons I stayed:
- Open AI treat their customers better (Better usage limits and resets; listen to user feedback and ship improvements often; Claude kept changing ToS)
@Technop54777070@ForwardEditor This used to work last week. Now it only spawns new threads with Sol/Terra. I have to switch to Luna xHigh manually. Hopefully, they'll fix it soon.
@thsottiaux ChatGPT Work is better. Vibe sounds unprofessional and hit- and-miss. Vibe coding has negative associations, IMHO. May as well call it ChatGPT Slop.
@andersonbcdefg Same here. Even when running Sol Medium as a planner and handing off implementation to Luna xHigh threads (all on standard speed), Codex runs much longer than before and burns a week's credit in a day. Considering switching back to GPT-5.5...
I'm very excited to be starting my new position as International Affairs Advisor to Prime Minister @netanyahu.
It is an honor and a privilege to be joining the Prime Minister as he leads us heroically through this critical time in the history of our nation and our state.