Share your thoughts, 1 month free Claude Pro on usSee more

Retrieval-Augmented Question Answering on DeepSearch Average

57.1SR

ActiveContext

Updated 3mo ago

Evaluation Results

Method	Links
ActiveContext 2026.04		57.1	6.6
Synapse 2026.04		56.9	80.7
Full Context 2026.04		53.9	46.7
Training-Free GRPO 2026.04		49	51.1
Gemini-3.0-flash 2026.04		42.1	7.8
ActiveContext 2026.04		41.5	7.3
Synapse 2026.04		41	43.6
Synapse 2026.04		40.7	85.9
Synapse 2026.04		39.1	41.9
Gemini-2.5-flash 2026.04		36.7	7.7
gpt-4o 2026.04		35	10.8
ContextCurator 2026.04		34.9	7.3
ActiveContext 2026.04		34.7	7.2
Mem0 2026.04		34.7	46.6
Training-Free GRPO 2026.04		33.7	35.7
Full Context 2026.04		33.4	34.4
Full Context 2026.04		33.4	34.4
AWM 2026.04		33.4	33.7
Training-Free GRPO 2026.04		33	33.6
gpt-4o-mini 2026.04		33	10.2
Full Context 2026.04		32.7	33.2
Qwen2.5-7B-Instruct 2026.04		30.4	9.9
Training-Free GRPO 2026.04		29.6	76.1