Codex users flag broken quota meter and GPT-6 Astra burning 5-hour limits in 10 minutes
A user letter cites ~40,000 Codex accounts on wrong usage limits, GPT-6 Astra draining a 5-hour quota in 10 minutes at 2.5x Opus 5.5 cost, and a 2,260-vote poll where 79.8% saw no workflow gain.
Dear @thsottiaux, Tonight you admitted ~40,000 Codex accounts were draining on the wrong usage limits. For weeks users said their quota was vanishing in minutes and got told they were imagining it. Now we know the meter was broken. Here is what your users are actually dealing with right now: 1. Your own power users warn people NOT to touch GPT-6 Astra for coding because it burns a 5-hour limit in 10 minutes while costing 2.5x Claude Opus 5.5. 2. In a 2,260-vote poll, 79.8% of users say daily Codex updates haven't improved their workflow or made it worse. The only update anyone feels is the reset button. 3. GPT-6 gets stuck in fix-test-review loops and timeouts, burning broken quota on retries. 4. Plus tier hits the wall after one task, and GPT-6 Luna loses to Haiku 5.5 at the exact same price. 5. Everyone is asking for 6.1 Astra and 6.1 Luna. The only model leaking in configs is 6.1 Terra. Stop handing out reset cards like party favors. Fix the meter, fix the loops, and ship 6.1 Astra.