MCR-Bench - JavaScript Defect Identification F1: leaderboard
Metric: F1 (%; matched defects). Source: arxiv.org. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Haiku 4.5 | 59.9 |
| 2 | GPT-5.2 | 57.3 |
| 3 | Gemini 3 Flash | 55.3 |
| 4 | DeepSeek V3.2 | 51.9 |
| 5 | GLM-4.7 | 48.9 |
| 6 | Qwen 3 Max | 42.2 |
| 7 | Kimi K2 | 39.1 |
Interactive version: theaggregate.ai/benchmark?slug=mcr-bench-javascript-defect-identification-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.