The AI arena is free today

Open Superagent

SWE Refactor Bench

Implementation

Progress Over Time

Interactive timeline showing model performance evolution on SWE Refactor Bench

State-of-the-art frontier
Open
Proprietary

SWE Refactor Bench Leaderboard

1 models
ContextCostLicense
1—1.0M$0.15 / $0.60
Notice missing or incorrect data?
About this benchmark

What is SWE Refactor Bench?

Repository-level software refactoring benchmark. MiniMax reports evaluations using the official setup with the Claude Code harness.

SWE Refactor Bench is a text benchmark evaluating models on agents and code tasks. LLM Stats tracks 1 models on this benchmark, scored on a 0–1 scale. The current average is 0.2, with the leader at 0.2.

Compare leaders on the best AI for agents and best AI for code leaderboards.

Current leaders

MiniMax M3.1 Flash from MiniMax currently leads the SWE Refactor Bench leaderboard with a score of 0.230 across 1 evaluated AI models.

1MiniMax M3.1 FlashMiniMax23.0%

FAQ

Common questions about the SWE Refactor Bench benchmark and leaderboard.

What is the SWE Refactor Bench benchmark?

Repository-level software refactoring benchmark. MiniMax reports evaluations using the official setup with the Claude Code harness.

What is the SWE Refactor Bench leaderboard?

The SWE Refactor Bench leaderboard ranks 1 AI models based on their performance on this benchmark. Currently, MiniMax M3.1 Flash by MiniMax leads with a score of 0.230. The average score across all models is 0.230.

What is the highest SWE Refactor Bench score?

The highest SWE Refactor Bench score is 0.230, achieved by MiniMax M3.1 Flash from MiniMax.

How many models are evaluated on SWE Refactor Bench?

1 models have been evaluated on the SWE Refactor Bench benchmark, with 0 verified results and 1 self-reported results.

Where can I find the SWE Refactor Bench dataset?

The SWE Refactor Bench dataset is available at https://github.com/Einsia/SWE-Refactor-Bench.

What categories does SWE Refactor Bench cover?

SWE Refactor Bench is categorized under agents and code. The benchmark evaluates text models.

Which model offers the best value on SWE Refactor Bench?

Among models scoring within 10% of the leader, MiniMax M3.1 Flash from MiniMax is the cheapest, at $0.15 per million input tokens with a score of 0.230.

How recent are the SWE Refactor Bench leaderboard results?

The SWE Refactor Bench leaderboard was last updated in November 2026 and currently includes 1 evaluated models.