Claude Opus 4.6 improves on Opus’ coding skills, and it now sustains agent tasks for longer and can run more reliably in larger codebases.
Claude Opus 4.6 is the latest version of the company’s most powerful class of AI models. Anthropic says that this release improves on Opus’ coding skills, and it now plans more carefully, sustains agent tasks for longer, can run more reliably in larger codebases, and has better code review and debugging skills, helping it to more effectively catch its own mistakes.
In addition to coding improvements, Claude Opus 4.6 also features improved abilities in tasks like running financial analyses, doing research, and using and creating documents, spreadsheets, and presentations.
It also features a 1M context window (beta), which is a first for Opus models, Anthropic said.
The latest version of OpenAI’s coding agent builds on the coding performance of the previous release as well as the reasoning and knowledge capabilities of GPT-5.2. It is capable of tackling long-running tasks involving research, tool use, and complex execution.
OpenAI also noted that GPT-5.3-Codex is the company’s first model that helped to create itself, as the team used early versions of it to troubleshoot training, manage deployment, and diagnose test results and evaluations.
It is now available through paid ChatGPT plans wherever Codex is enabled, including the app, CLI, IDE extension, and web, and API access will be enabled shortly.
The Developer Knowledge API allows developers to search and retrieve documentation for Google’s services in Markdown.
