List your company

Software Engineering Task Resolution on SWE-Bench Verified (with Token Usage)

61.6Success Rate (SR)Terminus 2
36.84843.27449.756.126Jun 11, 2026

Evaluation results

MethodLinks
Terminus 2Model=Qwen3.5-35B-A3B, Harness Category=Manual Harness2026.0661.61.47
HarnessBridgeModel=Qwen3.5-35B-A3B, Harness Category=Auto-Harness2026.0660.21.13
mini-SWE-agentModel=Qwen3.5-35B-A3B, Harness Category=Manual Harness2026.0659.85.92
Meta-HarnessModel=Qwen3.5-35B-A3B, Harness Category=Auto-Harness2026.0659.21.92
Qwen-CoderModel=Qwen3.5-35B-A3B, Harness Category=Manual Harness2026.0658.83.86
OpenHandsModel=Qwen3.5-35B-A3B, Harness Category=Manual Harness2026.0652.61.96
Meta-HarnessModel=GLM-4.7-Flash, Harness Category=Auto-Harness2026.0649.62.43
Terminus-KIRAModel=Qwen3.5-35B-A3B, Harness Category=Manual Harness2026.06469.77
HarnessBridgeModel=GLM-4.7-Flash, Harness Category=Auto-Harness2026.06461.48
mini-SWE-agentModel=GLM-4.7-Flash, Harness Category=Manual Harness2026.0645.43.91
Terminus 2Model=GLM-4.7-Flash, Harness Category=Manual Harness2026.0645.21.51
OpenHandsModel=GLM-4.7-Flash, Harness Category=Manual Harness2026.06423.69
Qwen-CoderModel=GLM-4.7-Flash, Harness Category=Manual Harness2026.06424.34
Terminus-KIRAModel=GLM-4.7-Flash, Harness Category=Manual Harness2026.0637.86.78