Simplify model selection in composer

## What changed On April 28, 2026 OpenAI moved model selection into the message composer for ChatGPT Plus, Pro, and Business users on web. The picker now sits where you type the prompt, and thinking effort controls for Thinking or Pro models appear inside the same menu.

The change removes the need to open a separate settings panel before each message. It applies only to the listed paid plans and remains unavailable to Free and Go users.

## Why it matters Frequent model switching becomes a single click instead of multiple steps. Vibe builders who test the same prompt across GPT-5.5 Instant, Thinking models, and Pro variants save seconds on every iteration.

The bet is that lower friction will increase usage of higher-cost models. It also signals OpenAI wants paid users to treat model choice as a routine part of prompt writing rather than a configuration task.

## How to use it Open ChatGPT on web while logged into a Plus, Pro, or Business account. Click the model name shown inside the composer box to open the picker, then select the model and any thinking effort level before sending.

No additional setup or feature flag is required. The option appears immediately for eligible accounts on the current web version.

## Watch for Confirm the change sticks by checking whether the old top-bar selector disappears entirely within the next month. Watch for mobile parity or Free-tier access as the next logical expansion. If OpenAI adds per-conversation defaults next, that would validate the direction toward persistent model preferences.