A clearer view of AI.
THE INDEPENDENT AI.

FRONTIER LAB / LLM · ESSENTIAL

OpenAI Developers (OpenAI Platform & Codex)

Official updates for developers building with Codex and the OpenAI API platform.

Visit developers.openai.com ↗@OpenAIDevs on X ↗

What OpenAI Developers (OpenAI Platform & Codex) does

Official updates for developers building with Codex and the OpenAI API platform. Covers the OpenAI API (models, Responses/Agents tooling) and Codex, OpenAI's coding agent available in ChatGPT, desktop, CLI and IDE. Status updates at status.openai.com.

Good for

  • Use Codex to automate coding, reviews and even finance/ops scripts
  • Build AI features on the OpenAI API
  • Follow model/pricing changes that affect your products

Try this in 10 minutes

  1. Write skill descriptions as triggers (when to use it), not summaries. (shown at 03:56 in the video)
  2. Tell review agents to make the smallest possible change, to stop scope creep. (shown at 05:28 in the video)
  3. Cap review loops (default 3 rounds; 5 for security/auth changes). (shown at 05:46 in the video)

From our fact-or-hype test list for “Theo (t3.gg): tstack, his agent skills for Claude Code and Codex (sol-loop, full-send, full-stop, takeover)”.

Honest take

No red flags in our notes so far

  • Fact or hype verdict on Theo (t3.gg): tstack, his agent skills for Claude Code and Codex (sol-loop, full-send, full-stop, takeover): a clean, real-world example of turning daily routines into named skills that call each other; worth copying the patterns (trigger-style descriptions, capped review loops, "when to ask me" rules) even if you never run his exact setup. Source video
  • Fact or hype verdict on Nidhi Singh: use Codex's computer use from Claude Code (or any MCP agent) on a Mac: a genuinely useful, honest setup guide for technical builders who want Claude Code to drive desktop apps. Not for beginners, and treat it as a hack that may break. Source video

Worth testing

Pricing

API pay-per-token; Codex included with ChatGPT plans (Free $0, Go $8, Plus $20, Pro $100-$500/mo, Business $20/user/mo yearly).

Prices as listed by the maker when we last checked (2026-10-29); usually in US dollars. Check the website for current prices.

Latest update

  • 2026-10-30: Agents API update: Auto Review of commands, async compaction, experimental time awareness, mid-session tool updates and automatic cross-session system-prompt caching.
  • : Agents API weekly update (30 Oct): Auto Review checks commands for bad behaviour before running (as in Codex), background (async) compaction, experimental time awareness so agents can act on asks like 'timebox CI fixes to 10 minutes', tools can be changed mid-session, and system prompts are now cached automatically across sessions.
  • : ChatGPT Pro and Business subscribers can link an API organisation and get monthly API credits (Pro $100/$200/$500; Business $20 per Standard and $100 per Premium seat, capped at $500/month per workspace); no rollover.
  • : Released GPT-6.1 Terra (gpt-6.1-terra) in the API and Codex: fast visual reasoning + computer use, $1/$5 per M tokens (50% below GPT-6.1 Sol), 1.05M context, supports skills, MCP and computer_use tools.
  • : Codex CLI 0.164.0 (23 Oct): /presets for up to 10 reusable prompts, inline /model and /permissions, image previews, enterprise MCP sign-in, credential masking; OpenAI says Codex CLI usage has doubled since DevDay.
  • : Decisions API is now generally available: apps and agents can classify text and images, evaluate inputs and decide the next step (which model, tool or action) without generating a full response; since beta it added native GPT-Live integration for voice apps that react as people speak, prompt caching and native JSON inputs.
  • : OpenAI frontier models and Codex agentic workflows are coming to Google Cloud's Model Garden, so enterprises can use them inside Google Cloud.
  • : Via its Patch the Planet initiative, OpenAI models found critical PHP engine flaws (PHP runs ~70% of websites), fixed with maintainers ahead of disclosure; every PHP site should upgrade. Use case: OpenAI packaged a Codex skill you point at its repo to check whether your website is affected.
  • : 12-hour prompt caching in the Responses API (GPT-5.6 and later): keep shared instructions, tools and context cached for half a day; 12h cache writes cost 2.5x input (vs 1.25x for the default 30 min), reads stay at the cached rate.
  • : Latest models much better at COBOL, Fortran, Visual Basic; GPT-6.1 Sol 86.6%-97.9% on OpenAI's COBOL migration eval (OpenAI's own benchmarks).

Is OpenAI Developers (OpenAI Platform & Codex) worth it? Quick answers

Is OpenAI Developers (OpenAI Platform & Codex) worth it?

OpenAI Developers (OpenAI Platform & Codex) is one of the tools we rate as essential to know about in the Frontier lab / LLM category. There is a free tier, so you can test it before paying.

How much does OpenAI Developers (OpenAI Platform & Codex) cost?

API pay-per-token; Codex included with ChatGPT plans (Free $0, Go $8, Plus $20, Pro $100-$500/mo, Business $20/user/mo yearly).

What is OpenAI Developers (OpenAI Platform & Codex) used for?

Official updates for developers building with Codex and the OpenAI API platform. Typical uses: use Codex to automate coding, reviews and even finance/ops scripts; build AI features on the OpenAI API; follow model/pricing changes that affect your products.