Rails gives coding agents the same thing it has always given developers: clear conventions, expressive code, and a complete framework for turning ideas into production software.
01
Convention Over Configuration
Rails gives agents a map. Standard names, folders, commands, and patterns help generated changes land closer to idiomatic Rails with less prompting.
02
Token Efficiency
Less code means more context. Ruby and Rails express product ideas with fewer tokens, helping agents make smaller edits and move faster from request to working feature.
03
Quality Training Data
The patterns are everywhere. Decades of public Rails code give models strong signals for controllers, models, views, tests, jobs, migrations, and the glue between them.
04
The One Person Framework
One person can ship more. Rails provides the full product stack. Add an AI coding agent, and a solo builder can take on broader work without stitching everything together from scratch.
Agents on Rails
Which model is best?
Two benchmarks, one harness. Atomic tasks ask whether a model knows Rails. Feature tickets ask whether it can ship, at default effort and at max effort.
20 feature-sized tickets on Fizzy, 37signals' kanban app, written the way a product person files them. Compare accuracy, speed, token efficiency, and cost.
Swipe to compare all model metrics →
Accuracy vs. token efficiency
The strongest models rise toward the top-right.
Swipe to explore every model →
Accuracy Higher is better
Mean tokens per run Lower is better
Further right uses fewer tokens
Methodology
Each model ran every ticket three times in September 2026 at the provider's default effort, and again at the highest effort the provider offers — 60 runs per model and effort, capped at 90 minutes, 400 steps, and $60 per run. Accuracy is the share of runs that passed both the app's own test suite and the ticket's hidden checks; differences of a few points between models are within run-to-run noise. Speed is the median run duration, and tokens and cost are means per run. Select any model or result for the underlying evaluation details.
21 small, self-contained tasks on Writebook, each turning on one Rails API. Compare accuracy, speed, token efficiency, cost, and API recall.
Swipe to compare all model metrics →
Accuracy vs. token efficiency
The strongest models rise toward the top-right.
Swipe to explore every model →
Accuracy Higher is better
Mean tokens per run Lower is better
Further right uses fewer tokens
Methodology
Each model ran every evaluation three times in August or September 2026, using the provider's default settings — 63 runs per model. Accuracy is the share of runs that passed the evaluation's hidden tests; refusals count as failures, and differences of a few points between models are within run-to-run noise. Speed is the median run duration, and tokens and cost are means per run. API recall is the percentage of runs in which the model reached directly for the target Rails API. Model-level medians come from run-level data, so they can differ slightly from the per-evaluation timings. GLM 5.3 ran on a coding-plan subscription, so it carries no dollar figures. Select any model or result for the underlying evaluation details.
Convention over configuration set the path for 20+ years of great training data for AI to use today. Not only does this mean agents do great with Rails, but also that squishy humans can quickly and confidently review the output without a jungle of distracting boilerplate.
Rails' convention over configuration explains why I got very good LLM output quality in the early days of LLM-assisted coding while many developers still thought it was unusable. Most Rails codebases look the same and on average is high quality.
One of the killer features of using Rails in this AI era is "Convention over configuration". If you keep your project close to the Rails defaults, the AI knows so much already about your project… It makes things 100x faster.