NITR (Agent): leaderboard
Metric: Agent-mode pass rate (%) over NITR's 21 C++ repository probes in nine maintainability dimensions: an edit passes only if the hidden functional tests and the structural maintainability oracle both pass; one run per probe; provider CLI agents (Codex CLI, Claude Code, Gemini CLI) with read-only repository access; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 57.14 | #60 |
| 2 | GPT-5.3 Codex | 57.14 | #68 |
| 3 | GPT-5 | 52.38 | #91 |
| 4 | GPT-5.2 Codex | 52.38 | #89 |
| 5 | Gemini 3.1 Pro (Preview) | 47.62 | #54 |
| 6 | GPT-5.4 | 47.62 | #76 |
| 7 | Claude Opus 4.5 | 47.62 | #79 |
| 8 | Claude Sonnet 4.6 | 42.86 | #85 |
| 9 | Claude Sonnet 4.5 | 38.1 | #138 |
| 10 | Gemini 2.5 Pro | 23.81 | #145 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=nitr-agent · How It Works · Data refreshed daily, snapshot 2026-10-11.