FeatureArc L4 - Stage F2P: leaderboard
FeatureArc tests feature development across successive stages in the same codebase. Results measure completed stages, complete features and uninterrupted progress. Stage fail-to-pass suite pass rate (%).
Metric: Stage fail-to-pass suite pass rate (%). Source: arxiv.org. Saturation forecast: Around March 2028. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 68.4 | #3 |
| 2 | GLM-5.3 | 56.3 | #42 |
| 3 | GPT-6 Astra | 55.5 | #5 |
| 4 | Grok 4.7 | 50.6 | #27 |
| 5 | Gemini 3.8 Flash | 49.4 | #25 |
| 6 | Kimi K3 | 42.2 | #35 |
| 7 | DeepSeek V4 Pro | 40.3 | #109 |
No result here: #7 GPT-6.1 Sol, #9 Gemini 4 Argon, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=featurearc-l4-stage-f2p · How It Works · Data refreshed daily, snapshot 2026-10-29.