DeepSeek has updated its flagship AI model, replacing the previous version after only 32 days. The new model, DeepSeek V2 Lite, offers significant improvements in cost and memory efficiency, being 77% cheaper and four times more memory-efficient than its predecessor. This rapid iteration suggests a focus on optimizing performance and accessibility for users. AI
IMPACT This rapid iteration and efficiency improvement in DeepSeek's models could accelerate adoption for teams prioritizing cost and memory.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →