Why did the token consumption rate for GPT 5.6 become so fast after GPT 6 went live? Before GPT 6, the same amount of work resulted in about 3% token consumption, but after September 5th, the token consumption for the same amount of work in GPT 5.6 Sol skyrocketed to 20%. Was this intentionally done by GPT? Please provide a reasonable explanation.
Same here, even after reset.
Same for me, even with writing simple methods, I reached my limit three times in three hours!
Are you using astra 6 on high at all? Im noticing that I can work for about an hour with astra 6 - medium. But the instant I put it on high it consumes my entire 5 hour limit in about 3 minutes before any substantial output.
It looks like (to me) that astra 6 high is a drastic jump from medium. They shouldve clarified that from the start, it consumes like its on ultra!
Maybe they just decided that high should be exclusive to the use-cases of 200 dollar plan guys only. Maybe this means ultra will actually feel like its worth using (if you can afford it.)