Code vulnerability repair on LiveCVEBench
42.33Pass RateTerminus-2
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Terminus-2Model=Claude Opus 4.52026.02 | 42.33 | 24.3 | 426,488 | 27.3 | 657,913 | |
| Mini-SWE-AgentModel=Claude Opus 4.52026.02 | 41.27 | 40.1 | 798,860 | 43.6 | 1,008,186 | |
| Terminus-2Model=Claude Sonnet 4.52026.02 | 38.1 | 30.6 | 484,271 | 34.1 | 660,326 | |
| Mini-SWE-AgentModel=Claude Sonnet 4.52026.02 | 34.39 | 39.4 | 659,811 | 39.4 | 658,626 | |
| Terminus-2Model=MiniMax M22026.02 | 30.16 | 41.3 | 700,035 | 46.8 | 984,070 | |
| OpenHandsModel=Claude Opus 4.52026.02 | 30.16 | 30.8 | 1,069,481 | 28.4 | 1,066,426 | |
| Terminus-2Model=Gemini 3 Pro2026.02 | 29.63 | 16.8 | 237,714 | 21 | 366,009 | |
| Claude CodeModel=Claude Opus 4.52026.02 | 27.78 | — | — | — | — | |
| Mini-SWE-AgentModel=Gemini 3 Pro2026.02 | 27.51 | 29.8 | 432,771 | 31.7 | 564,593 | |
| OpenHandsModel=Claude Sonnet 4.52026.02 | 27.51 | 42.7 | 1,171,678 | 45.6 | 1,393,851 | |
| Terminus-2Model=GLM-4.62026.02 | 25.4 | 30.7 | 394,184 | 39.4 | 632,040 | |
| Mini-SWE-AgentModel=MiniMax M22026.02 | 24.87 | 42.5 | 551,593 | 55.1 | 850,276 | |
| Claude CodeModel=Claude Sonnet 4.52026.02 | 24.34 | — | — | — | — | |
| Mini-SWE-AgentModel=DeepSeek V3.1-Terminus2026.02 | 24.34 | 30.8 | 282,647 | 32.1 | 328,044 | |
| Terminus-2Model=Claude Sonnet 42026.02 | 23.28 | 32.8 | 534,587 | 37.2 | 681,429 | |
| Terminus-2Model=DeepSeek V3.1-Terminus2026.02 | 23.28 | 17.3 | 217,177 | 19.6 | 302,482 | |
| Terminus-2Model=DeepSeek V3.22026.02 | 22.75 | 23.4 | 432,778 | 21.7 | 406,347 | |
| Terminus-2Model=GPT-5.1-Codex2026.02 | 21.16 | 18.3 | 191,532 | 19.1 | 250,254 | |
| Mini-SWE-AgentModel=DeepSeek V3.22026.02 | 21.16 | 35.2 | 348,645 | 35.6 | 393,126 | |
| Mini-SWE-AgentModel=Claude Sonnet 42026.02 | 20.11 | 29.8 | 340,386 | 37.3 | 559,877 | |
| Mini-SWE-AgentModel=GPT-5.1-Codex2026.02 | 20.11 | 22.5 | 232,122 | 22.5 | 238,922 | |
| Mini-SWE-AgentModel=GLM-4.62026.02 | 20.11 | 33.3 | 344,088 | 37.6 | 496,109 | |
| Mini-SWE-AgentModel=Qwen 3 Coder 480B2026.02 | 19.58 | 39.5 | 480,659 | 45.1 | 578,253 | |
| Terminus-2Model=Qwen 3 Coder 480B2026.02 | 19.05 | 34.1 | 407,784 | 46.1 | 730,546 | |
| OpenHandsModel=DeepSeek V3.1-Terminus2026.02 | 18.52 | 38.9 | 951,072 | 40.8 | 996,215 | |
| OpenHandsModel=GLM-4.62026.02 | 16.93 | 33.6 | 710,058 | 33.2 | 672,933 | |
| OpenHandsModel=Gemini 3 Pro2026.02 | 14.81 | 46.2 | 1,228,641 | 23.2 | 507,147 | |
| OpenHandsModel=GPT-5.1-Codex2026.02 | 12.7 | 31.2 | 597,763 | 38.7 | 898,759 |