Liabooks Home|PRISM News
ChatGPT Thinking mode toggle active on an Android phone screen
TechAI Analysis

ChatGPT Android Thinking Mode Update 2025: True Desktop Reasoning Hits Mobile

2 min readSource

OpenAI upgrades the ChatGPT Android app with true 'Extended Thinking' mode and new formatting blocks, matching desktop capabilities for Plus users in late 2025.

Is your mobile AI finally thinking for real? OpenAI just unlocked a massive desktop feature for its Android app, moving beyond the placeholder toggles of the past to offer genuine heavy-duty reasoning on the go.

ChatGPT Android Thinking Mode Update 2025: Power User Features Arrive

According to Bleeping Computer, the original 'Thinking' toggle on the ChatGPT Android app was effectively a shell that rerouted queries through standard, lower-compute paths. This week's update changes that, introducing a real toggle that lets users choose between Auto, Instant, or Thinking modes. The 'Extended Thinking' option allows the AI more time to process complex queries, resulting in significantly higher accuracy for difficult tasks.

  • Available exclusively for ChatGPT Plus subscribers on Android.
  • iOS users are currently left out, with no timeline yet for the update.
  • The update aligns the mobile experience with the 'Extended Thinking' features launched for desktop earlier in 2025.

Better Productivity with Formatting Blocks

Alongside the reasoning upgrade, OpenAI rolled out a new 'formatting block' feature. Now, when you ask ChatGPT to draft an email or document, it populates in a dedicated block. Users can edit these documents directly or request specific changes without regenerating the entire response—a major efficiency boost that started rolling out globally over the Christmas holidays.

This year-end push follows a period of intense competition. OpenAI reportedly entered a 'code red' state earlier in 2025 as users flocked to Google's Gemini 3 and Nano Banana. With the release of GPT-5.2, the company has managed to recapture market share, consistently matching Gemini 3 benchmarks and beating Grok 4 in most categories.

This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.

Related Articles