Description-guided molecule design on ChEBI-20 2022 (test) benchmark leaderboard · SOTA2 Research

Description-guided molecule design on ChEBI-20 2022 (test)

52.2Exact Match AccuracyBioT5+

Evaluation Results

Method	Links
BioT5+2024.02		52.2	0.872	12.776	0.907	0.835	0.779	0.353	0.579	100
BioT52024.02		41.3	0.867	15.097	0.886	0.801	0.734	0.43	0.576	100
BioT52023.10		41.3	0.867	15.097	0.886	0.801	0.734	0.43	0.576	100
MolT5-large2023.10		31.1	0.854	16.071	0.834	0.746	0.684	1.2	0.554	90.5
GPT-4-03142024.02		28	0.857	17.14	0.903	0.805	0.739	0.41	0.593	89.9
T5-large2023.10		27.9	0.854	16.721	0.823	0.731	0.67	1.22	0.552	90.2
MolXPT2024.02		21.5	—	—	0.859	0.757	0.667	0.45	0.578	98.3
MolXPT2023.10		21.5	—	—	0.859	0.757	0.667	0.45	0.578	98.3
MolFM-base2024.02		21	0.822	19.445	0.854	0.758	0.697	—	0.583	89.2
MoMu-base2024.02		18.3	0.815	20.52	0.847	0.737	0.678	—	0.58	86.3
GPT-3.5-turbo2024.02		13.9	0.79	24.91	0.847	0.708	0.624	0.57	0.571	88.7
GPT-3.5-turbo2023.10		13.9	0.79	24.91	0.847	0.708	0.624	0.57	0.571	88.7
MoT5-base2024.02		8.1	0.769	24.458	0.721	0.588	0.529	2.18	0.496	77.2
MolT5-base2023.10		8.1	0.769	24.458	0.721	0.588	0.529	2.18	0.496	77.2
MolT5-small2023.10		7.9	0.755	25.988	0.703	0.568	0.517	2.49	0.482	72.1
T5-base2024.02		6.9	0.762	24.95	0.731	0.605	0.545	2.48	0.499	66
T5-base2023.10		6.9	0.762	24.95	0.731	0.605	0.545	2.48	0.499	66
T5-small2023.10		6.4	0.741	27.703	0.704	0.578	0.525	2.89	0.479	60.8
GIT-Mol2024.02		5.1	0.756	26.315	0.738	0.582	0.519	—	—	92.8
Llama2-7B2024.02		2.2	0.693	36.77	0.808	0.717	0.609	4.9	0.149	76.1
GPT-3.5-turbo2024.02		1.9	0.489	52.13	0.705	0.462	0.367	2.05	0.479	80.2
GPT-3.5-turbo2023.10		1.9	0.489	52.13	0.705	0.462	0.367	2.05	0.479	80.2
RNN2023.10		0.5	0.652	38.09	0.591	0.4	0.362	4.55	0.409	54.2
Transformer2024.02		0	0.499	57.66	0.48	0.32	0.217	11.32	0.277	90.6
Llama2-7B2024.02		0	0.104	84.18	0.243	0.119	0.089	42.01	0.148	63.1
Transformer2023.10		0	0.499	57.66	0.48	0.32	0.217	11.32	0.277	90.6