List your company

Software Engineering Tool Use on SWE-bench Verified

76.2Success RateGemini-3-Pro
4.23222.91641.660.284Mar 10, 2026

Evaluation results

MethodLinks
Gemini-3-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0376.2
Claude-4-SonnetCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0372.7
Kimi-K2-0905Category=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0369.2
DeepSeek-V3.2-ExpCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0367.8
GPT-OSS-120BCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0362
Gemini-2.5-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=42026.0359.6
SWE-Dev-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.0319.5
DIVE-8B (RL)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0318.3
DIVE-8B (SFT)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0313.2
EnvScaler-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.0311.5
Qwen3-8B (base)Category=Ours, Temperature=1, Top-p=1, Number of runs=42026.0310.8
WebExplorer-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=42026.037