A-02 · Category market reportarchive-recorded · snapshot 2026-09-17
Research AI Tools: Category Market Report
Archive-derived market report for the Research category: 56 tool archives, median overall score 81.4/100 across five recorded dimensions, 46 free-tier listings and 166 recorded use-case entries. All figures are archive-recorded, computed from data/tools with the method and N published below.
Key numbers
Archives
56
data/tools · N=56
Overall median
81.4
score.overall · N=56
Free-tier listings
46
pricing.model · N=56
Median mo starting price
$10/mo
pricing.startingPrice.amount (unit=mo) · N=35
Recorded use-case entries
166
finder.useCases · N=56
Dated in snapshot window
55
updatedOn · N=56
Top 12 of 56 archives by overall score
score.overall as recorded; ties broken by slug
| # | Tool | Overall | Quality | Features | Usability | Performance | Value | Starting price | Updated |
|---|---|---|---|---|---|---|---|---|---|
| 1 | ChatGPT | 90.9 | 93 | 94 | 94 | 86 | 88 | $20/mo | 2026-09-06 |
| 2 | Claude | 88.9 | 94 | 88 | 92 | 86 | 84 | $17/mo | 2026-09-06 |
| 3 | Gemini | 88.5 | 89 | 88 | 91 | 85 | 90 | $19.99/mo | 2026-09-06 |
| 4 | Perplexity | 88.4 | 91 | 88 | 92 | 86 | 85 | $20/mo | 2026-09-06 |
| 5 | Semantic Scholar | 87.5 | 86 | 82 | 88 | 86 | 96 | $0/mo | 2026-09-15 |
| 6 | Poe | 87.1 | 86 | 90 | 88 | 84 | 88 | $19.99/mo | not recorded |
| 7 | OpenAlex | 87 | 86 | 84 | 82 | 86 | 96 | Free ($1/day API budget) · Pay-as-you-go $1 increments … | 2026-09-15 |
| 8 | DeepWiki | 85.4 | 85 | 76 | 90 | 83 | 94 | Free for public GitHub repositories · private repos via… | 2026-09-17 |
| 9 | NotebookLM | 85.2 | 88 | 78 | 88 | 82 | 90 | Free tier available · Plus via Google One AI Premium | 2026-09-06 |
| 10 | Elicit | 84.8 | 88 | 86 | 88 | 84 | 78 | $49/user | 2026-09-13 |
| 11 | DeepSeek | 84.7 | 86 | 80 | 80 | 84 | 92 | API: flash from $0.15/1M in (cache-hit $0.003) · v4-pro… | 2026-09-15 |
| 12 | Metaso | 84.5 | 82 | 78 | 88 | 84 | 92 | Free to use via metaso.cn (no public paid tiers listed) | 2026-09-17 |
Dimension medians
Median of each recorded dimension across the 56 Research archives (score.dims, scale 0–100).
Method
- Population: all archives in data/tools with category = "research" (N=56). No sampling.
- Medians use the standard median (middle element, or mean of the two middle elements rounded to 1 decimal) over score.overall and score.dims as recorded; scores are never recomputed or re-weighted.
- Starting price statistics use only numeric pricing.startingPrice.amount values in unit "mo" (N=35); archives with null amounts are excluded and displayed via pricing.label fallback instead.
- Benchmark coverage in this category: 0 of 56 archives carry a benchmark entry; T1 benchmarks recorded: 3 / 448 across the whole library.
- Freshness: 55 of 56 archives carry updatedOn inside the ten-day snapshot window 2026-09-06 → 2026-09-17.