List your company

Question Answering on ChronoTQA Tier 1 (147 Phenopackets-grounded onset questions)

86.4AccuracyDeepSeek V3
67.99272.77177.5582.329May 21, 2026

Evaluation results

MethodLinks
DeepSeek V3Retrieval Condition=NR2026.0586.4—
DeepSeek V3Retrieval Condition=ChronoMedKG2026.05851.4
DeepSeek V3Retrieval Condition=HPOA2026.0583.7—
Pooled (n=441)Retrieval Condition=ChronoMedKG2026.0582.12.3
Claude 3 HaikuRetrieval Condition=ChronoMedKG2026.0581.65.4
DeepSeek V3Retrieval Condition=PrimeKG2026.0581.6—
GPT-4o-miniRetrieval Condition=PrimeKG2026.0581.6—
GPT-4o-miniRetrieval Condition=HPOA2026.0581—
Pooled (n=441)Retrieval Condition=HPOA2026.0580.1—
Pooled (n=441)Retrieval Condition=NR2026.0579.8—
GPT-4o-miniRetrieval Condition=ChronoMedKG2026.0579.62.7
Pooled (n=441)Retrieval Condition=PrimeKG2026.0577.3—
GPT-4o-miniRetrieval Condition=NR2026.0576.9—
Claude 3 HaikuRetrieval Condition=NR2026.0576.2—
Claude 3 HaikuRetrieval Condition=HPOA2026.0575.5—
Claude 3 HaikuRetrieval Condition=PrimeKG2026.0568.7—