Share your thoughts, 1 month free Claude Pro on usSee more

Multimodal Question Answering on DesignQA

95Retrieval F1 (BoW)

GPT-5-MCERF-Hybrid

Updated 5d ago

Evaluation Results

Method	Links
GPT-5-MCERF-Hybrid 2026.01		95	55	-	-	-	-
GPT-5-MCERF-Main 2026.01		93	56	63	84	77	75
GPT-5-MCERF-HighReasoning 2026.01		92	51	64	85	80	81
GPT-4o-AllRules 2026.01		88	42	54	73	83	94
GPT-4-AllRules 2026.01		75	30	47	63	53	56
GPT-5-MCERF-SelfConsistency 2026.01		71	56	56	82	75	75
GPT-4o-MCERF-Main 2026.01		61	42	54	74	75	75
GPT-4o-RAG 2026.01		19	38	53	71	68	75
GPT-4-RAG 2026.01		18	36	42	53	30	54
Claude-Opus-RAG 2026.01		17	29	42	51	51	88
LLaVA-1.5-RAG 2026.01		11	28	39	48	41	44
Baseline Naive 2026.01		8	14	36	50	50	50
Gemini-1.0-RAG 2026.01		0	28	49	55	53	88
GPT-5-MCERF-Vision2Text 2026.01		-	-	63	81	82	94