MM-IssueLoc - File Localization: leaderboard
MM-IssueLoc tests whether repository-level localization systems use screenshots and other issue images to identify the files responsible for a software issue; scored by file Acc@5 with the best harness retained per underlying LLM.
Metric: File Acc@5 (%). Source: arxiv.org. Saturation forecast: Around December 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 38.96 | #110 |
| 2 | Claude Sonnet 4.6 | 37.42 | #90 |
| 3 | Kimi K2.5 | 35.74 | #144 |
| 4 | Gemini 3 Pro | 30.11 | #80 |
| 5 | Qwen 3.5 122B A10B | 25.7 | #173 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mm-issueloc-file-localization · How It Works · Data refreshed daily, snapshot 2026-10-20.