OpenAI has launched significant upgrades to ChatGPT Search, its AI-powered search feature that competes directly with Google’s AI search capabilities. The latest update brings substantial improvements in accuracy, reliability, and user experience.

## Enhanced Search Capabilities

The updated ChatGPT Search now delivers more precise results with better link summarization and improved follow-back functionality. OpenAI reports that the system generates fewer AI hallucinations—false or misleading information—resulting in higher-quality answers overall.

Key improvements include:
– **Better shopping detection**: The system now recognizes when users are looking for products and displays relevant results accordingly
– **Improved formatting**: Answers are structured for quick comprehension while maintaining detail and quality
– **Enhanced focus**: Results stay relevant to user queries without unnecessary information

## GPT-5 Thinking Model Gets Customizable Reasoning

Beyond search improvements, OpenAI has introduced a new toggle feature for GPT-5 Thinking that allows users to adjust reasoning effort levels. These levels, internally referred to as “juice” values, offer different processing intensities:

**Standard Users:**
– **Standard mode**: Level 18 (default balance of speed and intelligence)
– **Extended mode**: Level 64 (deeper analysis)

**Pro Users get additional options:**
– **Light mode**: Level 5 (fastest response time)
– **Heavy mode**: Level 200 (maximum reasoning capability)

## Gradual Rollout

These enhancements are being deployed gradually across OpenAI’s user base, ensuring system stability while expanding access to improved AI capabilities. The updates represent OpenAI’s continued effort to compete with major tech companies in the AI search and reasoning space.