Code Authorship Attribution on LLMAuthorBench benchmark leaderboard · SOTA2 Research

Code Authorship Attribution on LLMAuthorBench

98.92AccuracyCT5 ⊞ VIT

Evaluation Results

Method	Links
CT5 ⊞ VIT2026.03		98.92	93.82
CBE ⊞ VIT2026.03		98.81	94.11
CT5 ⊞ MXV2026.03		98.56	92.85
UXC ⊞ EFN2026.03		97.89	90.92
CT5 ⊕ VIT2026.03		97.12	92.39
CT5 ⊕ MXV2026.03		96.97	91.06
QWC ⊞ EFN2026.03		96.78	88.53
CBE ⊕ VIT2026.03		96.58	90.26
CBE ⊞ CNX2026.03		96.45	88.84
CT5 ⊗ VIT2026.03		96.35	90.04
UXC ⊕ EFN2026.03		96.31	89.73
CT5 ⊗ MXV2026.03		96.26	89.15
CBE ⊕ CNX2026.03		96.14	88.42
QWC ⊞ MXV2026.03		96.12	87.95
QWC ⊕ EFN2026.03		95.98	88.17
UXC ⊗ EFN2026.03		95.93	88.84
QWC ⊕ MXV2026.03		95.88	87.24
CBE ⊗ VIT2026.03		95.86	88.64
CT5 + VIT2026.03		95.62	87.88
CBE ⊗ CNX2026.03		95.58	87.02
CT5 + MXV2026.03		95.48	86.19
Orel et al.2026.03		95.4	—
CBE + VIT2026.03		95.28	86.47
UXC + EFN2026.03		95.18	87.45
QWC ⊗ MXV2026.03		95.12	86.92
QWC ⊗ EFN2026.03		95.08	87.73
CBE + CNX2026.03		94.85	86.42
QWC + EFN2026.03		94.63	85.36
QWC + MXV2026.03		94.36	85.14
CT52026.03		94.13	86.09
UXC2026.03		94.02	86.43
CBE2026.03		93.52	85.34
QWC2026.03		91.87	83.26
Gemini 3 Pro2026.03		83.71	69.94
GPT-5.12026.03		75.22	64.18
VIT2026.03		71.5	63.9
MXV2026.03		70.28	61.66
DeepSeek-V3.2-Exp2026.03		69.84	41.92
EFN2026.03		69.71	61.09
Gemini 3 Pro2026.03		66.45	56.38
CNX2026.03		65.09	57.21
GPT-5.12026.03		61.92	49.87
DeepSeek-V3.2-Exp2026.03		52.41	31.87