Clinical Summary Fact Scoring on Thoracic tumor board clinical cases benchmark leaderboard · SOTA2 Research

Clinical Summary Fact Scoring on Thoracic tumor board clinical cases

94.5Overall Fact Scoring AccuracyPhysician

Evaluation Results

Method	Links
Physician2026.04		94.5	95.5	100	100	98	88	97.7	94.7
Single-step2026.04		94.5	98.5	80.9	95.2	98	96.2	90.9	100
Physician2026.04		93.8	92.4	100	97.6	98	88	97.7	94.7
Multi-step2026.04		93.2	93.9	78.7	92.9	98	95.5	93.2	100
Single-step2026.04		91	98.5	80.9	66.7	95.9	95.5	90.9	100
Multi-step2026.04		91	93.9	78.7	71.4	98	95.5	93.2	100
Multi-Agent (Low-Autonomy)2026.04		90.8	93.9	80.9	95.2	98	88.7	86.4	100
Multi-Agent (Low-Autonomy)2026.04		87.5	92.4	80.9	69	98	88	86.4	100
Multi-Agent (High-Autonomy)2026.04		83.5	84.8	78.7	95.2	83.7	82.7	72.7	94.7
SecureGPT2026.04		83	89.4	66	88.1	91.8	84.2	70.5	89.5
Multi-Agent (High-Autonomy)2026.04		80	84.8	76.6	66.7	83.7	82	72.7	94.7
SecureGPT2026.04		78.2	87.9	63.8	50	89.8	84.2	70.5	89.5
Most Recent Oncology Note2026.04		65.2	74.2	61.7	69	75.5	58.6	54.5	78.9
Most Recent Oncology Note2026.04		62.5	72.7	61.7	47.6	75.5	57.9	54.5	78.9