OpenAI Codex Guide: Code Generation & Practical Tips

Conversations and code on personal ChatGPT and Codex plans, including Pro, are used for training by default. One ChatGPT setting also covers new Codex tasks, but environment training has a separate setting. The guide also covers what not to enter, what to do if you submit API keys or customer data, and the exceptions for deletion and review.

In the ChatGPT desktop app (Codex), right-click the floating pet and choose Hide Mini to remove it quickly. No official documentation describes a setting that disables pets in the app, so if a pet reappears, check the display shortcut first. This guide compares the steps, Mini, and the CLI, web, and IDE extension differences, using official documentation and a user-provided Windows screenshot.

The Codex message "Selected model is at capacity" is caused by congestion on OpenAI's side. Official incident records and our own execution logs both classify it as server overload. This guide covers the steps to take in order (wait, check incidents, try another model), how it differs from usage limits, 401, and thread not found, and what to record if it happens again.

ChatGPT dot is a continuous AI agent that keeps working on tasks between conversations, included at no extra charge with Pro. In our to-do app test, dot handled generating the app and adding undo, but screen checks were completed in Codex. The article covers pricing, usage limits, memory and permissions, delegation to Codex, and how to stop work, so you can decide how to split work between dot and Codex.

ChatGPT Pro comes in three tiers: 100, 200 and 500. Since October 29, 2026, the Work and Codex allowances are 10, 20 and 50 times Plus (calculated by this site), and the official monthly prices are $100, $200 and $500, so two 100 subscriptions and one 200 cost the same and give the same total allowance; local-currency prices vary by country or region. The article also explains how speed settings use up allowance, how and when to use banked reset passes, and how they differ from credits.

If you use GPT-6 Sol in Codex, switching to GPT-6.1 Sol is fine. Released on September 29, 2026, it is recommended for complex coding and long-running work, with the same context limit and standard API prices, half-price cached input, and a knowledge cutoff 10 days newer. The guide covers switching steps in Codex, the difference between new and existing chats, Plus and Pro usage limits, and how to compare both at Medium effort before you decide.

Even with the same Pro subscription, image generation draws from different usage pools in regular ChatGPT Chat versus Codex and Work. Claude models do not directly create photos or illustrations, but Claude Code can produce images through HTML, SVG, code, or external image APIs. The article diagrams each entry point and walks through a practical method: set the display size first, then specify the composition, separate reference material from new content, and select from a few outputs by checking them at full size and at the actual display size.

Opening your sessions one by one to switch the model and effort every time a new model comes out is a chore. When I tried it in my own setup on September 24, 2026, asking the assistant to "change them all to this" worked in both Claude Code Desktop and Codex, without me touching each session. In Claude Code Desktop, the tools the app gives the assistant changed the sessions without sending a message, and the changes survived an app restart. However, it only sees local desktop sessions, the default is the 20 most recent, and it can't change the session making the request. In Codex, the default for new chats was changed in config.toml, and the assistant used internal machinery to change the 115 existing chats in my setup one by one. Chats that were open in the app each received one confirmation message. This article covers how the two differ, what they can and can't reach, example prompts, and 4 checks: keep the old values, define the scope, re-read everything, and check one with your own eyes.

OpenAI released GPT-6 Sol and GPT-6 Luna on September 22, 2026 (US time). Sol is priced at 1/5 of Astra ($2 input / $10 output) and Luna at 1/100 ($0.10 input / $0.50 output), and both are available in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu, and in the API. Free and Go get Luna only, in the desktop app. Neither model was in regular Chat at launch, but since October 7 GPT-6 in Chat runs on Sol for Plus, Pro, Business and Enterprise, and since October 8 on Luna for Free and Go. GPT-5.6 remains available during the transition period. Written for people who currently use Astra at low (shown as Light in the app), this article separates its sources by publisher: the official use-case guidance (Sol at Medium, Luna at High, Astra from Light) and how it differs from the app's default preset (Sol Light); OpenAI's AutomationBench result where Sol xhigh beat Astra low; the values in Artificial Analysis's comparisons (42% for Astra low vs 19% for Sol medium on Terminal-Bench 4.0, but 59% vs 58% on business workflows); and AI Arte's small-scale test (every setting passed the short tasks, which is no basis for a ranking). It also explains why unit-price ratios don't carry over directly to the total cost of a job or to how much a subscription covers, and where the decision turns: Sol medium for everyday work, Sol low or Luna high for short, easy-to-verify jobs, and Astra for hard design work and important reviews.

When Codex shows your history but can't send a message with "thread not found," the conversation may not be lost. This article explains the difference between saved history and the conversation that accepts new instructions, and shares a case where sending recovered after reloading. It then covers the order to try fixes, how to hand off the task without deleting history if the problem persists, and what to include in a report.

A newer model is smarter but more expensive, or so the thinking usually goes. Yet comparing GPT-6 Astra with GPT-5.6 Sol on unit price alone settles nothing, because Astra finishes the same job on fewer tokens. In measurements by the independent benchmarking firm Artificial Analysis, running Astra at its weakest setting, low, costs almost exactly what Sol costs at high for one task ($0.82 against $0.81) while scoring higher (46 against 42), on one third the output tokens and a quarter of the wait before the reply starts. Working from the figures as of September 18, 2026, this article covers cost, tokens and speed at each level of reasoning effort, the breakdown that explains why a 2.5x unit price still lands on the same total, what the numbers look like once Sol's promotional price ends, the way the gap in price narrows on coding agent work, and the steps for measuring all of it on your own work.

OpenAI released GPT-6 Astra on September 3, 2026. The API is generally available with no gating, the model ID is gpt-6-astra, the context window is 1,050,000 tokens, pricing is $10 in / $50 out per million tokens, and the knowledge cutoff is April 30, 2026. But "it launched" and "your plan can use it" are two different things: on ChatGPT Plus it does not appear in the regular chat interface, and the way in is ChatGPT Work and Codex. This article covers availability by plan, how to think about pricing when a 2x unit price can still be cheaper per task, the benchmarks OpenAI actually named, the cybersecurity capability that reached "Critical" under the Preparedness Framework for the first time and the line OpenAI drew around it, the four things that break when you migrate (sampling parameters removed, Responses API, no none reasoning effort, the renamed cache setting), and a comparison with Claude Opus 5, Fable 5.1, and Gemini 3.8 Flash — restricted throughout to what could be confirmed against primary sources. It also shows, with each publisher named, how the same DeepSWE v1.1 gives Claude Opus 5 73.7% in the OpenAI table but 68.8% in the Anthropic announcement.