Proactive Demand Detection on IntentFlow (Demand) benchmark leaderboard · SOTA2 Research

Proactive Demand Detection on IntentFlow (Demand)

85.5Business MetricsGemini-3-Flash

Evaluation Results

Method	Links
Gemini-3-Flash2026.04		85.5	90.3	86.4	78.8	83.2	89.7	88.4	76.1	89.9	65	83.3
IntentFlow2026.04		80.6	87.2	86.9	79.4	86.1	88.1	86.5	83.6	86.4	77	84.2
GPT-5-Nano2026.04		80.5	87.4	88.4	77.2	58	53.4	51.6	77.5	74.4	63.1	71.2
GPT-5-Mini2026.04		79.9	80.8	78.5	71.3	78.8	75.8	80.6	77.3	80.5	68.5	77.2
IntentFlow2026.04		79.2	86	85.4	78.8	85.3	86.2	86.3	82.4	85.7	75.8	83.1
GPT-oss-120b2026.04		75.2	75.8	69.2	66.5	68.8	72.6	68.6	68.9	69.3	67.8	70.3
Gemini-3-Flash2026.04		74.7	85.1	85.7	73	84	88	84.8	77.9	85.2	69.9	80.8
GPT-5-Mini2026.04		71.1	69.1	73.2	65.8	78.3	60.3	69.5	63.4	67.3	46.5	66.5
Claude-Haiku-4.52026.04		70.5	75.5	72.8	66.7	58.6	68.3	57.5	63.9	68	59.9	66.2
GPT-oss-120b2026.04		68.6	62.2	61.1	58	56.6	66.4	50.5	49.3	52.4	45.9	57.1
GPT-5-Nano2026.04		66.6	78.8	77.1	63.9	70.2	71.4	69.7	76.1	74.8	66.3	71.5
Qwen3.5-Flash2026.04		63.7	63	58.1	56.5	61	63.5	62.9	60.9	61.5	59.8	61.1
Claude-Haiku-4.52026.04		62.9	57.2	55.6	45.1	21.7	39.7	15.8	32.4	36.9	21.7	38.9
DeepSeek-V3.22026.04		60.5	66.4	62.2	60.1	53.5	57.5	54.9	71	71.2	58.6	61.6
Qwen3-30B-A3B2026.04		52.3	60.8	64.7	57	57.9	60.6	52.7	59.5	70.6	52.5	58.9
Gemini-2.5-Flash-Lite2026.04		49.4	55.5	48	51	52.8	52.8	51.9	59.1	55.2	45.1	52.1
DeepSeek-V3.22026.04		44.7	42.4	33.8	34.7	7	18.1	12.6	53.5	54.2	21.7	32.3
Qwen3.5-Flash2026.04		34.6	32	23.2	20.2	30.8	31.9	29.5	29.6	28.6	30.6	29.1
Gemini-2.5-Flash-Lite2026.04		17.6	30.2	15.2	17.1	15.4	18.1	11.6	27.5	28.6	6.4	18.8
Qwen3-30B-A3B2026.04		15.1	29.1	51.5	27.5	19.6	43.1	8.4	24.6	53.6	15.9	28.8