DeepSeek’s V4-Flash undercuts rivals on running costs while matching mid-tier benchmark scores
Artificial Analysis, a San Francisco research firm, found DeepSeek’s newly released V4-Flash model to be substantially less expensive to run on benchmark tests than widely known AI models, with an estimated average cost of $0.03 per test. V4-Flash charges $0.14 per million input tokens and $0.28 per million output tokens, and scored 50 on an Intell…