DeepSeek has released its V3.1 model, which has undergone continued pre-training on 840 billion tokens. This new model reportedly outperforms Anthropic's Claude 4 Sonnet while operating at a significantly lower cost, approximately 11% of Sonnet's price. The release highlights advancements in efficient large language model training and performance. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
RANK_REASON Frontier-lab model release with system card.