Testing the Best Deep Research Tools for Academic Research

This document provides an overview of testing experiences with various deep research tools for academic purposes. The evaluation focuses on the tools' effectiveness in sourcing, formatting, and handling research information.

Objectives of Testing

• Information Retrieval: Testing if tools can fetch recent information (past two years).
• References and Citations: Quality and quantity of references provided.
• Explanations: Clarity and usefulness of information presented.
• Multimedia Support: Availability and incorporation of images, tables, and graphs.
• Exportability: Ease of using the information in other formats or applications.

Evaluation of Tools

1. Chat GPT

• Pros:
  • Solid multimedia integration (figures, tables).
  • Clear presentation of recent breakthroughs and key developments.
• Cons:
  • Discrepancy in the number of cited references (claimed vs. shown).
  • Limited academic exportability.
• Score: 4 (rounded from 3.5)

2. SciSpace

• Pros:
  • Strong focus on academic references (up to 200+, exportable).
  • Specific focus on recent papers.
• Cons:
  • Less comprehensive text information.
  • Format is slightly less clear.
• Score: 3

3. Perplexity

• Pros:
  • Interactive references with clickable links to sources.
  • Good filtering for peer-reviewed content.
• Cons:
  • Limited multimedia support.
  • Inconsistent reference count (9 vs. 49).
• Score: 3

4. Gemini

• Pros:
  • Extensive reference listing (up to 253).
  • Exportable to DOC and highly detailed.
• Cons:
  • Lack of multimedia integration.
• Score: 4

5. Manis AI

• Pros:
  • Separate and focused documents for various research aspects.
  • Exportable in PDF with useful segmentation.
• Cons:
  • References not directly linked in the text.
• Score: 4

6. Storm (Stanford)

• Pros:
  • Completely free and informative.
• Cons:
  • Poor reference interface; requires extra effort to verify source recency.
  • Lack of multimedia and export features.
• Score: 2

Conclusion

• Top Performers: Gemini and Manis AI offer comprehensive references and export features, making them suitable for academic research.
• Recommendation: Choose Gemini for detailed referencing and Manis AI for a structured approach. SciSpace remains a solid option for purely reference-focused tasks.

These tools vary in terms of specific strengths, allowing researchers to select based on particular needs such as reference management or detailed analysis.