Breaking

Tight prompts keep GPT-6 Astra Plus jobs to 1% of weekly quota

September 13, 2026 at 16:45 EDT

Foundation Models
Software Dev & Coding
AI Agents

A tightly scoped GPT-6 Astra run on ChatGPT Plus used about 1% of the weekly Work and Codex allowance; a harder follow-up finished in 9.5 minutes at 6%. Developers comparing notes say the difference is not the model’s speed so much as whether the prompt forbids extra work.

September 2026 · OpenAI GPT-6 Astra

Keep the prompt tight. Keep the meter low.

On ChatGPT Plus, a scoped GPT-6 Astra job used about 1% of the weekly Work and Codex allowance. A harder follow-up took 9.5 minutes and 6%. Unbounded runs still empty the week in a single session.

Weekly allowance used (Plus)

Constrained run

Harder follow-up
9.5 min

Six stacked blocks versus one: the follow-up still stayed in single-digit weekly spend.

Weekly used, tight prompt

Follow-up on Plus

5–45

Plus Astra msgs / 5h (est.)

Test cap in the recipe

Constraint recipe

1 Clean prompt, necessary context only

→

2 No scope expansion

→

3 Fail twice → new strategy

→

4 10 tests, then stop and report

What works

Hard limits in the prompt can keep Plus Astra jobs in the low single digits of weekly usage.

What still burns quota

Open-ended sessions still drain the week: one ~30-minute Codex run left 8% weekly; Pro 20x heavy use left 40% in under a day.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.