SWE-Cycle - FullCycle Solve Rate: leaderboard

SWE-Cycle FullCycle Solve Rate is the percent of the 489 SWE-Cycle (2026) issues that a code agent resolves perfectly end to end, with full marks for environment setup, implementation and tests on both static review and execution. Higher is better.

Metric: FullCycle solve rate (%): share of issues whose FullCycle score is perfect (every phase full marks on both static and dynamic checks by SWE-Judge (Claude Opus 4.5 agent) static code review and dynamic execution, each scored 0, 1 or 2), on the 489 SWE-Cycle issues (225 SWE-bench Verified, 203 SWE-bench Pro, 61 SWE-bench Multilingual, filtered for contamination, lifecycle complexity and reliable tests); one session covering environment setup, implementation and tests (3 hours); agents run in the OpenCode harness in isolated Docker containers; higher is better. Source: arxiv.org. Saturation forecast: Around February 2028. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GLM-5.113.5#198
2Claude Sonnet 4.612.27#173
3GPT-5.410.84#135
4SWE-Cycle Qwen-3.5 (checkpoint unspecified)6.75
5MiniMax-M2.74.29#539
6Kimi K2.52.15#266

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=swe-cycle-fullcycle-solve-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.