Open data: prompts and picks
Everything is free to use under CC BY 4.0: credit Agent Stack Picks with a link. The latest issue is October 2026.
Files
- picks-2026-10.csvShares per category and tool, with 95% intervals and per-context counts
- picks-2026-10-raw.csvEvery parsed pick: model, category, context, repeat, tool
- picks-2026-10.jsonThe issue file: all picks, model metadata, totals
- prompts.jsonEvery prompt, the system prompt, the schema
Always the newest: picks.csv and picks.json redirect to the latest issue. Past issues stay at their own addresses.
Fields of the picks CSV
| Field | Type | Meaning |
|---|---|---|
| issue | text | Month of the run, for example 2026-10. |
| category | text | Category slug, for example auth. |
| tool_id / tool_name | text | Canonical tool after alias normalisation. Names not in the alias table appear as other:<slug>. |
| picks / category_picks / share | number | How often the tool was named, all parsed picks in the category, and picks divided by category picks. |
| ci95_low / ci95_high | number | 95 percent Wilson interval of the share. |
| picks_nextjs, picks_fastapi, picks_vibe, picks_eu / share_eu | number | Picks per context, and the share within the EU context. |
| models_naming_it | number | Number of different models that named the tool at least once. |
License
CC BY 4.0. Use, share and adapt, including commercially, with credit and a link. We publish counts of tool names, not the models' free text.
Agent Stack Picks (October 2026). What AI models pick for a new app: monthly index, issue 2026-10. https://agentstackpicks.com/2026/10. CC BY 4.0.
Changelog
- 2026-10-19: Prompt set 2026.1 published: 24 categories, 4 contexts. First measurement (issue 2026-10) with open models (Gemma 4 31B, Nemotron 3 Super 120B, Laguna S 2.1) and chat models (Claude Opus 5.5, Claude Sonnet 5.5, Gemini 3.1 Pro, Gemini 3.6 Flash, GPT-5.6, Kimi K3); more models and repeats follow.
Changes to the prompt set, the alias table, categories or models are listed here and on the methodology page.