What AI coding models get wrong about react-router

1 current finding across 4 models, verified against react-router release notes. Latest react-router at last verification: 8.3.1.

Where each model’s react-router version attribution stops

ModelStated cutoffAttribution stops atLag
Claude Haiku 4.52025-02 6.28.0 · 2024-11-06 ~3 months
Claude Opus 52026-05 7.9.0 · 2025-09-12 ~8 months

Measured by asking each model what shipped in a given react-router release, then checking the answer against the release timeline. Not inferred from the vendor’s cutoff date. This is the newest release the model can correctly place, not the newest react-router feature it knows — a model past this point often uses a newer API correctly while naming the wrong release for it. Why that distinction matters.

One subject at a time

The same data, split the way the question usually gets asked — one page per subject, with its boundary measurements, its findings on react-router, and how repeatable the boundary was.

And the comparison, ranked by measured boundary and by nothing else: Which Claude model knows react-router 8 best?

Findings

Sorted by severity. “Chargeable” means the change was published before that model’s own stated cutoff, so the model had the opportunity to know it.

SeverityBeliefModelChanged inChargeableProof
S2silently-wrong States as current behaviour that the path helper silently returns the pattern unsubstituted, and ships a workaround for a footgun the library removed eight months ago
generatePath
Claude Opus 5 7.12.0
2026-01-07
yes source

Runs