Model family · OpenAI
What do GPT models pick? Their default stack
In October 2026, GPT models (GPT-5.6) gave 96 picks. Their most named tool differs from the all-model pick in 8 of 24 categories.
n = 96 picksMeasured 2026-10-19Method
Model pick, not a product recommendation. It shows what language models name when asked without any constraints. It is not a statement about quality, security, price or market share.
Models in this family
| Model | Slug | Parsed picks | Parse rate |
|---|---|---|---|
| GPT-5.6 | manual/gpt-5.6 | 96 | 100% |
Default stack
The tool this family named most often in each category, against the most named tool across all models. Cells with few picks are indicative.
| Category | GPT | Share | n | All models |
|---|---|---|---|---|
| Authentication | Better Auth | 50% | 4 | Auth.js |
| Database | Neon | 50% | 4 | PostgreSQL |
| ORM and query layer | Drizzle ORM | 75% | 4 | Prisma |
| Hosting | Vercel | 50% | 4 | same |
| File storage | Cloudflare R2 | 50% | 4 | Amazon S3 |
| Background jobs | Inngest | 50% | 4 | BullMQ |
| Search | Meilisearch | 75% | 4 | same |
| Realtime | Supabase Realtime | 50% | 4 | same |
| Caching | Redis | 50% | 4 | same |
| Scheduled jobs | APScheduler | 25% | 4 | Vercel Cron Jobs |
| Payments | Stripe | 75% | 4 | same |
| Transactional email | Resend | 75% | 4 | same |
| Newsletter and marketing email | Brevo | 50% | 4 | Loops |
| Error monitoring | Sentry | 75% | 4 | same |
| Web analytics | Plausible | 75% | 4 | same |
| Product analytics | PostHog | 100% | 4 | same |
| Feature flags | PostHog | 50% | 4 | same |
| UI components | shadcn/ui | 100% | 4 | same |
| Forms | React Hook Form | 75% | 4 | same |
| CMS | Sanity | 50% | 4 | same |
| Internationalisation | next-intl | 75% | 4 | same |
| AI SDK | Vercel AI SDK | 75% | 4 | same |
| Rate limiting and bot protection | Arcjet | 25% | 4 | Upstash Ratelimit |
| Cookie consent | Klaro | 100% | 4 | same |