Public data release · 2026-08-15 · schema v1

AI Prompt Trends Research: 3,239 Published Examples

This release describes the published PicGens corpus as of 2026-08-15. It counts public rows by the stored model label, existing category tags and the month of the source post. It is a transparent corpus snapshot—not a claim about search demand, rankings, engagement, market share or model quality.

Published rows

3,239

Rows where is_published = true was observed in the read-only run.

Model labels

10

“Unlabelled model” stays visible; no model is inferred from prompt wording.

Latest ingest

2026-08-15

Latest represented source date: 2026-08-14.

Published rows by stored model label

GPT Image 2 is the largest labelled group in this snapshot, with 1,636 rows. These are corpus counts only: they do not measure which model is best, most popular in search or more likely to reproduce a result.

Model label counts in the 2026-08-15 PicGens snapshot
Stored keyLabelPublished rows
gpt-image-2GPT Image 21,636
nanobanana-proNano Banana Pro468
unknownUnlabelled model455
nanobanana-2Nano Banana 2285
midjourneyMidjourney116
nanobananaNano Banana94
seedreamSeedream85
seedanceSeedance60
gpt-image-1GPT Image 139
veoVeo1

Published rows by existing category tag

Category tags describe the existing editorial labels attached to rows. A row can carry more than one tag, so category totals intentionally do not add up to the corpus total.

Portraits

1,960

Posters & visuals

1,440

Illustration & 3D

1,141

Ads & product

792

Brand & logo

315

Video

274

Wallpaper

152

Published rows by source-post month

The month dimension uses the stored source-post timestamp, not the ingest date. It helps describe when the represented public examples were created; it is not a time-series measure of demand or traffic.

Source-post month counts in the 2026-08-15 PicGens snapshot
MonthRows
2023-101
2025-043
2025-051
2025-074
2025-081
2025-093
2025-101
2025-117
2025-1240
2026-01142
2026-02113
2026-0359
2026-0487
2026-05185
2026-061,264
2026-071,218
2026-08110

Method

Read published rows only and aggregate by model, existing category tag and tweet-created month; the 2026-08-15 run read 3,239 rows through a paginated Supabase query.

  • Keep “unknown” as a visible bucket instead of inferring a model from prompt wording.
  • Do not expose raw prompt contents, row identifiers, creator handles, private fields or per-row behaviour.
  • Counts are a dated corpus snapshot; they are not search volume, ranking, engagement or market share.

Copy rate: unavailable — no copy event export was authorised for this snapshot.

Limitations

  • The latest ingested row is 2026-08-15 and the latest tweet-created date represented is 2026-08-14; this is not a live trend feed.
  • Category tags can overlap, so category totals are not expected to sum to the corpus total.
  • No GSC, GA4, CrUX, backlink or copy-rate data was available or inferred.

No raw prompt text, row IDs, creator handles, source URLs, private fields or per-row behaviour are distributed with this release.

Machine-readable access

Agents and researchers can use the JSON or CSV distribution as a stable, dated citation target. The files describe aggregate dimensions and limitations; they do not expose individual creator content or grant rights to reuse any source image or prompt.

For prompt workflows, see the prompt variable analyzer and image ratio recommender. For the live gallery, browse model collections and use-case collections.