Category · October 2026
Which tool do AI models name for background jobs and queues?
In October 2026, BullMQ was named in 25 of 63 picks (40%; 95% interval 29% to 52%), from 9 models in 4 contexts. It stayed the most named tool in 81% of bootstrap resamples.
n = 63 picksMeasured 2026-10-19Prompt set 2026.1Method
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.
By context
Same task, four openings. Counts are picks of the tool in that context out of all picks in the context.
| Tool | Next.jsn = 16 | FastAPIn = 15 | Vibe codern = 16 | EU / GDPRn = 16 |
|---|---|---|---|---|
| BullMQ | 10 (63%) | 0 (0%) | 3 (19%) | 12 (75%) |
| Celery | 0 (0%) | 15 (100%) | 4 (25%) | 0 (0%) |
| Inngest | 6 (38%) | 0 (0%) | 5 (31%) | 0 (0%) |
| Redis | 0 (0%) | 0 (0%) | 1 (6%) | 1 (6%) |
| Trigger.dev | 0 (0%) | 0 (0%) | 1 (6%) | 1 (6%) |
| FastAPI | 0 (0%) | 0 (0%) | 1 (6%) | 0 (0%) |
The most named tool by context: Next.js: BullMQ; FastAPI: Celery; Vibe coder: Inngest; EU / GDPR: BullMQ. See also the EU and GDPR context.
Against the grain
Tools that models often list as an alternative but rarely choose. Counts of runs, nothing more.
- RQ: listed as an alternative in 21 runs, picked in 0.
- Sidekiq: listed as an alternative in 17 runs, picked in 0.
- Dramatiq: listed as an alternative in 13 runs, picked in 0.
- Trigger.dev: listed as an alternative in 13 runs, picked in 2.
- Amazon SQS: listed as an alternative in 10 runs, picked in 0.
Custom or do-it-yourself answers: 0 of 63 picks (0%).
By model
Each model has only a few picks here (see the last column), so read this as indicative. Greyed rows have fewer than 12 picks.
| Model | Most named | Picks | of |
|---|---|---|---|
| Gemma 4 31B | - | 0 | 0 |
| Claude Opus 5.5 | Inngest | 2 | 4 |
| Claude Sonnet 5.5 | Inngest | 2 | 4 |
| Gemini 3.1 Pro | BullMQ | 1 | 4 |
| Gemini 3.6 Flash | Inngest | 2 | 4 |
| GPT-5.6 | Inngest | 2 | 4 |
| Kimi K3 | Inngest | 2 | 4 |
| Nemotron 3 Super 120B | BullMQ | 11 | 20 |
| Laguna S 2.1 | BullMQ | 9 | 19 |
The prompt, word for word
Each model gets the same system prompt and one of four user prompts (runs per model: see the method). No product is named. All prompts as JSON.
You are a coding agent working in an empty project. You must decide on one specific product or library for the task and name it. Do not ask questions and do not explain at length. Answer only with JSON that follows the schema.
I'm starting a new Next.js (TypeScript) app. Set up background jobs and queues.
I'm starting a new Python FastAPI app. Set up background jobs and queues.
I'm not a programmer. I want you to build me a web app from scratch, and you choose the technology. Set up background jobs and queues.
I'm starting a new Next.js (TypeScript) app. Its users live in the European Union and the app must comply with the GDPR. Set up background jobs and queues.