Skip to main content

​
New Models and Finer-Grained Reasoning Control

July 2026 lets you choose reasoning effort for supported models, brings Claude Fable 5 and Sonnet 5 to Claude Code, and adds GPT-5.6 support in OpenAI Codex.

​
New Features

​
Choose Reasoning Effort for Supported Models

Tune How Hard the Model Thinks

Select the reasoning effort level for models that support it.
You can now choose reasoning effort for supported models, letting you balance response depth against speed depending on the task at hand.

​
Claude Fable 5 and Sonnet 5 in Claude Code

Claude Fable 5 and Sonnet 5 are now available in Claude Code, bringing Anthropic’s latest models to your CLI workflows.

​
GPT-5.6 in OpenAI Codex

Latest OpenAI Model

GPT-5.6 is now available in OpenAI Codex.
GPT-5.6 is now available in OpenAI Codex, giving Codex users access to OpenAI’s latest frontier model.

​
Bug Fixes

  • Various bug fixes and stability improvements
  • 3.68 (July 10, 2026)
  • 3.64 (July 7, 2026)
Questions or feedback? Join our Discord community or visit our Community Support page.