Measuring Codex Usage Burn Rate: Astra (high) Without Subagents
Melvin Vivas · X post · 2026-09-10 · Open on X
Topics: AI Dev Tools & Productivity, LLMOps, Deployment & Monitoring · Level: intermediate
Summary
A quick, hands-on cost test of Codex with the Astra model at the 'high' reasoning level and no subagents, covering a review, plan and code workflow. Over 17 minutes the weekly limit dropped from 64% to 62%. The post works out the burn rate and how long a full weekly limit would last. It also shows a simple way to estimate the cost of your own coding-agent setup.
Key points
- Test setup: Codex with Astra at high reasoning, no subagents, doing a review → plan → code workflow.
- Weekly usage went from 64% to 62% between 12:37 and 12:54, so 2% in 17 minutes.
- That is about 0.118% per minute, or roughly 7.1% per hour.
- At that rate a full 100% weekly limit lasts about 850 minutes (around 14h 10m) of active work.
- Next test: Astra at xhigh reasoning, which is reported to be cheaper.
- Method: note the usage % at a start and end time, then extrapolate to estimate how long your limits will last.
Resources mentioned
- OpenAI Codex · tool · openai.com · paid · recommended by both Bashiri Smith & Melvin Vivas
OpenAI's coding agent. In the diagram it writes code, fixes review findings and drives the build loop. The creator also used it to make this video.
Also in: Developers moving from ChatGPT/Codex to Claude (Melvin Vivas on X · notes), Creator Opinion: Codex Is Still the Best at Computer Use (Melvin Vivas on X · notes), OpenAI Codex Ambassador applications are open (Melvin Vivas on X · notes), GPT-Live-1.1: Multilingual Voice Update for ChatGPT and Codex (Melvin Vivas on X · notes) and 264 more - GPT-6 Astra · tool · openai.com · paid
The model announced in the quoted launch post, pitched as the developer's most capable model for work, coding, science and cybersecurity, and able to operate a computer.
Also in: Glimpse: A Visual Reasoning Model That Uses a Visual Scratch Space (Melvin Vivas on X · notes), Is AI Getting Cheaper? How to Measure LLM Cost Per Task and Per Intelligence Level (Melvin Vivas on X · notes), GitHub Copilot Adds Intelligent Local Model Routing (Project HydraFusion) (Melvin Vivas on X · notes), Comparing Opus 5.5, GPT-6 Astra and Grok 4.7 on OpenRouter (Melvin Vivas on X · notes) and 55 more
Try this
- Track the usage % at a start and end time to calculate your own burn rate.
- Compare reasoning levels (high vs xhigh) to see which is cheaper for your workflow.
More in AI Dev Tools & Productivity
- Codex Astra Making a Demo Video with HyperFrames
- Measure Your Own Codex Token Burn: Orchestrator+Subagents vs Single Agent
- Single-Agent Codex Astra xhigh Usage for a Feature Build
- Editing a GitHub Repo From ChatGPT Work on iOS
- Codex as an All-Round Dev Companion
- Analyzing Codex Traces with jsonl-viewer