Scientific Text Generation on VTechAGP CELLS 1.0 (test) benchmark leaderboard · SOTA2 Research

Scientific Text Generation on VTechAGP CELLS 1.0 (test)

14.59ROUGE2Phi-4

Evaluation Results

Method	Links
Phi-42025.05		14.59	4.08	78.87	34.15
Sci-LoRA2025.05		14.02	3.99	79.1	44.36
Qwen2.52025.05		11.92	3.07	78.95	43.43
Mixtral2025.05		9.83	2.55	79.46	49.35
OPT2025.05		9.77	—	74.74	34.05
GPT-3.52025.05		9.54	2.56	80.23	41.29
Mistral2025.05		8.98	1.51	77.84	51.06
GPT-4o2025.05		8.7	2.15	79.97	41.9
LLaMA32025.05		7.97	1.84	79.17	49.55