Agentic Question Answering on AirQA benchmark leaderboard · SOTA2 Research

Agentic Question Answering on AirQA

51.85SGL ScoreGemini-2.5-Pro

Evaluation Results

Method	Links
Gemini-2.5-Pro2025.09		51.85	18.58	67.01	40.49	51.53	29.58	29.95	33.86	53.28	46.55	38.23	44.14
GPT-4o2025.09		45.58	10.53	52.13	35.56	39.61	23	25.6	33.86	47.54	38.76	29.09	35.96
Claude-3.7-Sonnet2025.09		45.3	15.17	58.68	27.46	43.96	22.07	24.64	27.56	44.26	39.32	29.64	36.52
DeepSeek-R12025.09		41.03	11.46	41.67	22.54	35.1	15.96	20.77	20.47	39.34	30.4	26.59	29.29
Qwen2.5-72B-Instruct2025.09		39.03	10.84	55.21	37.32	41.71	13.15	28.02	30.71	45.9	37.74	28.53	35.07
o1-mini2025.09		37.04	12.07	45.14	24.65	35.43	14.55	22.22	22.83	36.07	31.07	26.04	29.61
Llama-3.3-70B-Instruct2025.09		29.06	9.29	42.71	24.3	32.37	8.92	21.74	19.69	30.33	28.47	19.94	26