TRAIL SWE — leaderboard

TRAIL (Trace Reasoning and Agentic Issue Localization) SWE subset: evaluates LLMs on trace-based debugging and issue localization in software engineering contexts.

Metric: Joint Accuracy. Source: huggingface.co. Status: years away from saturation. 5 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro (Preview 05-06)5
2GPT-4.10
3Llama 4 Scout Instruct0
4Gemini 2.5 Flash (Preview 04-17)0
5Llama 4 Maverick Instruct0

Interactive version: theaggregate.ai/benchmark?slug=trail-swe · How the rankings work · Data refreshed daily, snapshot 2026-07-22.