A benchmark suite for evaluating how coding models solve real React Native tasks.
Groups map to top-level folders under evals/.
| Group | Path | Status |
|---|---|---|
| animation | evals/animation |
Active |
| async-state | evals/async-state |
Active |
| navigation | evals/navigation |
Active |
| react-native-apis | evals/react-native-apis |
Active |
| expo-sdk | evals/expo-sdk |
WIP |
| brownfield | evals/brownfield |
WIP |
| nitro-modules | evals/nitro-modules |
WIP |
| lists | evals/lists |
Active |
Want a group that is not listed here? Open an issue to request it. Contributions are also welcome.
bun install
bun runner/run.ts --model openai/gpt-4.1-mini --output generated/my-generated
bun runner/judge.ts --model openai/gpt-5.3-codex --input generated/my-generatedFor full command reference and workflows, see docs and CONTRIBUTING.md.
Methodology and scoring details are documented in the benchmark methodology whitepaper.
The benchmark evaluates model-generated React Native implementations using requirement-based assessment. Each eval specifies a fixed task context and a set of explicit, judgeable requirements. Model outputs are judged against these requirements using file-level evidence, and per-eval scores are computed from requirement outcomes with optional weighting. Aggregate run metrics summarize performance across evals under a consistent evaluation protocol.
If you want to request new features to be evaluated, open an issue. We are open to covering the most popular ecosystem libraries and will continue expanding coverage.
Contributions are welcome. Start with CONTRIBUTING.md and AGENTS.md.
MIT (LICENSE)