<ul><li>DeepSeek-V3, the general-purpose AI model, achieves the highest score among non-reasoning models.</li><li>The updated model, DeepSeek V3-0324, ranks first in benchmarks on the Artificial Analysis platform.</li><li>It surpasses competing models in the GPQA Diamond benchmark, achieving a score of 66%.</li><li>DeepSeek plans to release the upcoming DeepSeek-R2 reasoning model based on V3-0324.</li></ul>

DeepSeek-V3 is the Highest Scoring Non-Reasoning Model – ‘A Milestone for Open Source’

Discover more