OpenAI Expands Deep Research to ChatGPT Plus as AI-Powered Knowledge Assistants Gain Traction

OpenAI has expanded its Deep Research feature to ChatGPT Plus users, making its AI-powered research assistant more accessible beyond its original $200-per-month Pro tier.

The update introduces embedded images with citations and improved handling of uploaded documents, enhancing how the AI integrates user-provided files into structured reports.

These refinements come as OpenAI faces growing competition from Google Gemini, Perplexity AI, and latest with xAI’s Grok 3, each developing their own AI-driven research solutions.

The move signals OpenAI’s intent to establish Deep Research as a broader tool for professional investigations, offering AI-assisted multi-step reports rather than single-response chatbot answers.

However, citation accuracy remains a concern, and OpenAI has yet to confirm how it plans to resolve lingering issues with references aligning incorrectly.

How Deep Research Works and What’s Changing

Deep Research operates differently from ChatGPT’s standard instant responses. Instead of generating quick answers, it compiles structured reports over five to thirty minutes, retrieving and analyzing multiple sources before formatting them into a detailed output.

The most noticeable upgrade in this expansion is the inclusion of embedded images with citations, addressing user concerns about having to manually verify sources.

Previously, Deep Research generated text-only reports, requiring additional effort to confirm information. The AI now integrates visual references directly into its reports, offering more transparency in how it sources its findings.

Another key improvement is the AI’s ability to process uploaded files more effectively. Users can submit PDFs, spreadsheets, and other documents, which Deep Research incorporates into its findings.

This is particularly useful for professionals dealing with proprietary reports, enabling AI-assisted analysis that blends internal documents with publicly available data.

Despite these enhancements, OpenAI acknowledges that citation accuracy still needs refinement. “Since the initial launch, we’ve made some improvements to deep research: embedded images with citations in the output and better at understanding and referencing uploaded files,” the company stated.

However, it has not provided details on when or how it will improve the consistency of AI-generated citations.

Competition Intensifies as Google, Perplexity, and xAI Expand AI Research

The AI research space is becoming increasingly competitive, with major players developing alternative tools that take different approaches to knowledge retrieval.

Google added Deep Research to its Gemini AI platform, first launching on desktop before expanding to Android. Unlike OpenAI’s single-pass retrieval model, Gemini’s AI applies an agent-based system, refining its search multiple times before compiling a final structured response. This iterative approach allows Gemini to adjust its outputs dynamically, making it more adaptable for complex research queries.

A few days ago, Google introduced its AI-powered co-scientist research assistant, designed to generate scientific hypotheses. The system evaluates vast amounts of scientific literature, detects gaps in knowledge, and ranks potential areas for further study.

Perplexity AI has positioned itself as another alternative, launching its own Deep Research feature on February 15. Instead of relying on pre-trained models, Perplexity continuously scans the web for the latest information, providing responses that incorporate up-to-date sources. While this method offers more recent data, it also raises concerns about misinformation, as AI-generated reports pulling from live web sources may lack rigorous fact-checking.

Elon Musk’s artificial intelligence company xAI recently released its much improved Grok 3 model with an similar feature called deep search. For real-time research, Deep Search gives Grok 3 an edge over many models, but its accuracy issues put it behind OpenAI’s Deep Research and Perplexity AI.

Hugging Face has taken a different approach with its Open DeepResearch initiative, launched on February 6. Unlike OpenAI and Google, which operate proprietary AI models, Hugging Face provides an open-source framework that allows developers to customize and modify AI-driven research workflows. By offering a transparent alternative, it challenges the subscription-based models dominating the market.

Subscription-Based AI Research: The New Business Model

OpenAI’s decision to bring Deep Research to ChatGPT Plus aligns with a broader shift in AI monetization, where advanced features are being moved behind subscription tiers. Originally positioned as a high-end feature for Pro users, the tool’s availability at a lower price suggests OpenAI is exploring ways to expand adoption while still maintaining paid access to AI research capabilities.

This model is not unique to OpenAI. X (formerly Twitter) has also shifted AI services behind paywalls, doubling the cost of its Premium+ subscription to $40 per month while restricting access to Grok 3. The company has further integrated AI into business applications, launching Grok AI-generated ad tools to automate campaign creation for advertisers.

Meanwhile, Google continues to expand its AI ecosystem. The recent introduction of Gemini 2.0 Pro Experimental enhances Gemini AI’s ability to a two-million-token context window, allowing it to handle vast amounts of information simultaneously.

Table: AI Model Benchmarks – LLM Leaderboard 

Last Updated on March 3, 2025 11:27 am CET