Misinformation Detection on PHEME benchmark leaderboard · SOTA2 Research

Misinformation Detection on PHEME

69.2AccuracyMetaAdapt

Evaluation Results

Method	Links
MetaAdapt2026.01		69.2	62.5	50.8	43.3
gpt-5-mini2026.01		58.6	54.9	55.4	54.7
gpt-5-mini2026.01		58.4	55.3	56	55
gpt-4.12026.01		57.9	52.4	52.5	52.4
IT_ra2026.01		57.8	57.8	59.1	56.2
Qwen-8b2026.01		57.1	36.1	36.1	35.8
gpt-4.12026.01		57	51.7	51.8	51.6
RAEmoLLM2026.01		57	52.7	53	52.5
Qwen-14b2026.01		56.7	37.5	38	36.5
Qwen-32b2026.01		56	54.8	55.6	53.7
MOSE2026.01		55.5	54.6	55.4	53.4
RAAR-14b2026.01		54.1	57.2	58.2	53.5
RAAR-8b2026.01		53.6	55.8	56.7	52.8
Qwen-14b2026.01		53.5	54	54.7	52
Qwen-32b2026.01		53.5	53.1	53.6	51.5
Deepseek-R2026.01		53.1	52.2	52.6	50.9
Qwen-8b2026.01		52.6	35.6	35.9	34.1
Deepseek-C2026.01		51.8	51.8	52.1	50
Deepseek-R2026.01		51.2	51.5	51.8	49.5
Deepseek-C2026.01		49.6	50.3	50.4	48.1
CANMD2026.01		48.2	57	56.9	48.1
Llama-8b2026.01		46.2	34.4	31.5	30.9
IT_zs2026.01		46.1	56.8	56.1	45.9
Llama-8b2026.01		44.8	33.9	33.9	29.8
MDFEND2026.01		43.3	62.2	57.5	42
EDDFN2026.01		33.1	53.6	50.7	27.8