DeepSeek Releases Low-Cost V4-Flash AI Model Ahead of IPO
DeepSeek launched V4-Flash, the world's least expensive known AI model to operate, while preparing for an initial public offering and developing a more powerful version.
DeepSeek, a Chinese AI startup, released its V4-Flash model on Friday, establishing it as the least expensive well-known AI model to operate globally. According to research from Artificial Analysis, V4-Flash costs an average of 3 cents per test, making it more than 100 times cheaper to run than Anthropic's Claude Fable 5.
Other industry benchmarks show a significant cost gap, with Moonshot AI's Kimi K3 costing 86 cents and OpenAI's GPT-5.6 Sol costing $1.86 per test. However, the cost-efficiency comes with a trade-off in performance. V4-Flash scored 50 out of 100 on the Artificial Analysis Intelligence Index, trailing behind the models from OpenAI, Anthropic, and Moonshot AI.
The release comes as DeepSeek prepares for an initial public offering and develops a more powerful V4-Pro model. Meanwhile, Alibaba Group entered the scaling market on Monday by unveiling its largest AI model to date, Qwen3.8-Max.
Keep reading in the app
The full story and every source, free in the app.
Coming soonGoogle Play