DeepSeek AI: Cheapest to Run, Outperforms Rivals

A version of Chinese startup DeepSeek’s flagship AI model is the least expensive to run on benchmark tests globally. It is more than 100 times cheaper to run than Anthropic’s Claude Fable 5. DeepSeek officially released its V4-Flash model last week. This model charges $0.14 per million input tokens and $0.28 per million output tokens. Research firm Artificial Analysis estimates the V4-Flash’s average cost at 3 cents per test. This compares to 86 cents for Kimi K3 from Moonshot AI, $1.86 for OpenAI’s GPT-5.6 Sol, and $3.15 for Claude Fable 5. What's interesting is that this cost comparison accounts for the amount of data a model processes to complete a task. The V4-Flash model scored 50 out of 100 on Artificial Analysis’s Intelligence Index. This is the same score as Google’s Gemini 3.6 Flash. The bottom line is that DeepSeek is offering ultra-low-cost AI alternatives, which could impact businesses looking to deploy AI at scale.