米OpenAIは3月5日(日本時間)、「GPT‑5.4」を発表した。同社でもっとも高性能かつ効率的な業務用途向けフロンティアモデルだ。
Databricks has released KARL, an RL-trained RAG agent that it says handles all six enterprise search categories at 33% lower cost than frontier models.
Databricks' KARL agent uses reinforcement learning to generalize across six enterprise search behaviors — the problem that breaks most RAG pipelines.