SID AI has launched SID-2, a new agentic search model. This model utilizes reinforcement learning for training. A single agent of SID-2 achieved an NDCG score of 0.585 in just over 3 seconds, while a swarm of 12 agents reached an NDCG score of 0.720, surpassing GPT-6. AI
IMPACT Sets a new benchmark for agentic search models, potentially influencing future AI development in information retrieval.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →