OpenAI has begun beta testing composer predictions in Codex, a feature that suggests a user’s next message based on the conversation and communication style. The announcement was published on October 9, 2026, and applies to Pro users.
Beta access and controls
Composer predictions are available in the Codex desktop app for local tasks. A suggested message can be accepted with the Tab key, edited, and then sent. OpenAI also provides a setting to disable the feature.
OpenAI Developers described composer predictions as “one of the most loved new features” tested internally, a characterization that remains the company’s own assessment rather than an independently measured result.
The rollout drew comparisons with Claude Code. One reply characterized the feature as something Claude Code had offered “for forever,” while another claimed Anthropic’s product had supported it for more than a year. A separate reply questioned why the feature was limited to Pro users rather than being available to everyone.
The announcement’s accompanying comparison-themed graphic includes the terms “CLAUDE CODE” and “CODEX.”
Source: OpenAI Developers on X
