工具条件未记录 / high / mini-swe-agent · Self-computed with mini-swe-agent harness; uses highest scoring Gemini thinking level reported on the main leaderboard.
DeepSWE · v1.1 · Score · 工具条件未记录 / high / mini-swe-agent · Self-computed with mini-swe-agent harness; uses highest scoring Gemini thinking level reported on the main leaderboard.
Regular display scope. Rankings are grouped by metric and evaluation protocol. Recorded tool access, budgets, and frameworks are shown separately. Scores remain available for reference when conditions are unspecified.
Score %
Tool access not recorded / max / Framework not recorded · DeepSWE v1.1; 113 long-horizon software engineering tasks; average of five trials.
DeepSWE · v1.1 · Score · Tool access not recorded / max / Framework not recorded · DeepSWE v1.1; 113 long-horizon software engineering tasks; average of five trials.
Tool access not recorded / high / mini-swe-agent · Self-computed with mini-swe-agent harness; uses highest scoring Gemini thinking level reported on the main leaderboard.
DeepSWE · v1.1 · Score · Tool access not recorded / high / mini-swe-agent · Self-computed with mini-swe-agent harness; uses highest scoring Gemini thinking level reported on the main leaderboard.
Tool access not recorded / max / Framework not recorded · DeepSWE v1.1, 113 long-horizon software-engineering tasks; five-trial average.
DeepSWE · v1.1 · Score · Tool access not recorded / max / Framework not recorded · DeepSWE v1.1, 113 long-horizon software-engineering tasks; five-trial average.