Clinical Text-to-SQL on EpiAskKB 2025 (test) benchmark leaderboard · SOTA2 Research

Clinical Text-to-SQL on EpiAskKB 2025 (test)

93.3AccuracyGemini3Pro

Evaluation Results

Method	Links
Gemini3Pro2026.02		93.3	0.708	0.623	0.604	3.8
Gemini3Pro2026.02		93.3	0.62	0.532	0.715	20.6
Gemini3Pro2026.02		88.8	0.642	0.578	0.593	5.9
Gemini3Pro2026.02		88.8	0.635	0.542	0.595	10.7
KimiK22026.02		78.9	0.599	0.418	0.853	27.8
Claude4.5S2026.02		78.9	0.593	0.395	0.611	0
Claude4.5S2026.02		78.9	0.575	0.426	0.622	16.8
Claude4.5S2026.02		78	0.633	0.4	0.744	30.8
Claude4.5S2026.02		77.6	0.595	0.466	0.694	10.3
KimiK22026.02		76.1	0.632	0.446	0.815	41.7
GPT-OSS2026.02		75.5	0.615	0.41	0.828	34.3
KimiK22026.02		75.2	0.635	0.444	0.791	44.7
Pooled2026.02		74.4	0.6	0.461	0.762	33.6
KimiK22026.02		74.3	0.548	0.417	0.817	40.5
GPT-OSS2026.02		71.6	0.6	0.416	0.731	31.2
GPT-OSS2026.02		71.6	0.615	0.384	0.831	34.2
GPT-OSS2026.02		71.6	0.612	0.391	0.818	35
Qwen32026.02		55	0.618	0.505	0.818	59.2
Qwen32026.02		54.2	0.617	0.531	0.748	42.6
Qwen32026.02		54.1	0.63	0.498	0.786	65.8
Qwen32026.02		53.3	0.605	0.544	0.764	54.3