OpenAI Codex

Evidence (18)

• 6The New Stack·10/5/2026product_launch
  OpenAI’s Codex sprint starts with a speed boost

  OpenAI has begun a 28-day Codex shipping sprint by making GPT-6 Astra and GPT-6.1 Sol faster, with reported generation speeds rising to about 50 tokens per second. The move comes alongside OpenAI’s ongoing usage reset strategy, while developers debate whether daily improvements or periodic resets matter more than higher limits.

• 6The New Stack·9/17/2026research
  Study Links AI Coding Tools to Overuse and Reduced Understanding

  A survey in the AI Coding Addiction Report suggests that frequent use of AI coding tools can foster dependency-like behavior, with many developers skipping breaks and working longer hours. The report also says heavy AI users are more likely to receive rewards from managers even when they do not fully understand the code they ship.

• 5Hacker News·9/14/2026security
  Lessons from migrating long preprompts to self-hosted Ollama

  The post discusses practical issues encountered while moving large preprompt workflows from Anthropic Claude Opus to a self-hosted Ollama setup. It argues that hosting inference on your own hardware can better protect sensitive session metadata and prompts from third-party providers. The write-up also criticizes frontier AI vendors’ security posture and retention practices.

• 6The New Stack·9/12/2026research
  OpenAI hires Git AI founders to improve Codex ROI tracking

  OpenAI has hired Git AI founders Aidan Cunniffe and Sasha Varlamov to join its Codex team and help businesses better measure how coding agents perform and what returns they generate. The move suggests OpenAI plans to use Git AI’s open-source tracking technology to show where Codex adds value across individual and team workflows.

• 8The New Stack·9/9/2026security
  OpenAI lets AI block engineer code merges

  OpenAI has made automated security review mandatory for engineer pull requests, and its AI models can prevent a merge if they detect a vulnerability. The company also uses models for code review, regression detection, dependency upgrades, and maintenance tasks, with some internal benchmarks described as “superhuman” in code review.

• 7The New Stack·9/8/2026regulation
  Anthropic faces suit over unclear Claude Max weekly limits

  Anthropic’s Claude Max plans promise up to 20x more usage than Pro, but developers can still hit a separate weekly cap, according to an expanded class-action lawsuit. The case argues that Anthropic did not clearly disclose these limits, highlighting a broader challenge in pricing AI coding subscriptions with unpredictable compute usage.

• 9Hacker News·9/8/2026research
  OpenAI and researchers clash over AI-assisted Navier–Stokes proof

  NYU professor Tristan Buckmaster and Anthropic mathematician Levent Alpöge announced partial progress on the Navier–Stokes existence and smoothness problem using AI models, but said OpenAI may have learned of their approach before publishing its own proof attempt. OpenAI later posted a proof claim based on work from an unreleased next-generation model, alongside a timeline suggesting its effort began after rumors of progress on Millennium Prize problems.

• 7Hacker News·9/6/2026research
  AI-assisted proof of the Spherical Hadwiger Theorem

  A preprint by Wang and Wu claims to prove the long-open Spherical Hadwiger Conjecture with AI assistance, specifically noting OpenAI Codex helped develop proof details, find gaps, and prepare the manuscript. The post frames this as another sign that AI is starting to affect research mathematics, even in niche areas with limited economic value and sparse training data.

• 4Hacker News·9/3/2026product_launch
  Codex service outage

  This Hacker News item links to a GitHub issue reporting an outage affecting OpenAI Codex. The post appears to be a status-style incident report rather than a product update, with limited public detail in the thread.

• 3OpenAI Blog·8/26/2026product_launch
  loveholidays uses Codex to broaden software building

  OpenAI highlights how loveholidays is using Codex to make software development more accessible across the company. The approach helps non-engineering teams turn ideas into products more quickly and supports faster delivery of internal and customer-facing work.

• 6Hacker News·8/21/2026security
  Codex AWS Bedrock Bug Triggers 10x Charges

  A GitHub issue reports that OpenAI Codex users on AWS Bedrock experienced unexpectedly high billing, with costs reportedly multiplying by 10x. The discussion centers on a charging bug and its impact on users running Codex through Bedrock, rather than a new product or model release.

• 6The New Stack·8/19/2026framework_update
  Codex tests asynchronous developer messaging while continuing work

  OpenAI is testing an async messaging tool for Codex that lets the agent ask a developer a question, keep working on unrelated tasks, and incorporate the reply later. The change, merged into the public Codex repository, replaces blocking input flow with a non-blocking message acceptance model, though the feature is not yet broadly announced or tied to a specific release target.

• 7OpenAI Blog·8/18/2026product_launch
  Asana cuts years of testing work to weeks with Codex

  Asana said it used OpenAI Codex to replace an outdated testing system in two weeks. The company said the effort would have taken about five years and roughly $12,000 if done through the prior approach.

• 7Hacker News·8/13/2026product_launch
  Netlify Adds OpenRouter Models to AI Gateway and Agent Runners

  Netlify has partnered with OpenRouter to let projects use a wider range of models through its AI Gateway, while also expanding the coding models available in Agent Runners. The company says it tested identical prompts across multiple models to compare output quality, functionality, and credit efficiency for common site-building tasks. The report highlights newer open models such as Kimi K3, GLM 5.2, and DeepSeek V4, along with Netlify’s own AXIS evaluation tooling.

• 4Hacker News·7/28/2026security
  Codex Security

  This Hacker News item points to OpenAI's Codex Security repository, which appears to focus on security-related information for Codex. The discussion suggests interest in how OpenAI is documenting or handling security considerations around its coding tool.

• 4Hacker News·7/19/2026framework_update
  OpenAI cuts Codex model context window from 372k to 272k

  OpenAI updated Codex to use a smaller model context size, reducing the window from 372k to 272k tokens. The change appears in a GitHub pull request and has drawn discussion on Hacker News about the tradeoff between context length, cost, and performance.

• 4Hacker News·7/18/2026product_launch
  OpenAI Codex Usage Limits Reset Tracker

  This Hacker News post tracks repeated resets of OpenAI Codex usage limits, apparently announced ad hoc by OpenAI staff on X. The log shows frequent resets tied to user milestones, reliability issues, and infrastructure scaling as Codex and ChatGPT Work usage grows.

• 3Hacker News·7/15/2026open_source
  Brainless releases shadcn components styled like Claude Code, Codex, and Grok

  The post showcases Brainless, a set of shadcn components that mimic the interfaces of Claude Code, OpenAI Codex, and Grok. The examples are presented as reusable UI blocks for adding agent-style experiences to web apps, including pricing and command-driven workflows. The item is primarily a UI/component release rather than a new model or platform.