Urgent.News

600+ sources. One page. See who else covered it.

Editions

AI

DeepSeek launches V4 Pro at prices up to 14 times higher than V4 Flash

DeepSeek launches V4 Pro at prices up to 14 times higher than V4 Flash

Chinese AI startup DeepSeek unveiled its latest model, V4 Pro, on Thursday, signaling a significant price increase compared to its more affordable V4 Flash model. The V4-Pro-0813 costs $1.32 per million input tokens and $3.96 per million output tokens, which is roughly 9 times and 14 times the respective prices of V4 Flash. The V4-Pro-0813 boasts enhanced capabilities for AI agents, accessible via the company's API, app, and web products.

DeepSeek's decision to price the V4-Pro-0813 so much higher than its predecessor comes after the model's unexpected success in benchmark tests. Released last month, V4 Flash had outperformed a preview version of V4 Pro, leading to an assumption that the official launch represented a substantial improvement in technology. Independent evaluations by Artificial Analysis have reinforced this perception, assigning V4 Pro a score of 53 on its Intelligence Index, compared to 40 for V4 Flash.

The index examines various features, including agentic work tasks, tool use, coding, scientific reasoning, and long-context performance.

The pricing strategy positions DeepSeek's V4 Pro as a premium offering in the market, reflecting the rapid advancements the company has made in AI technology. This move aims to regain its competitive edge in the highly competitive AI landscape, which has seen other Chinese companies like Moonshot AI, Zhipu AI, MiniMax, Alibaba, and ByteDance make significant strides.

DeepSeek had previously faced challenges in sustaining its momentum, including a reported plan to launch a new fundraising round at a valuation of $74 billion, following a significant investment of $7.4 billion in June. The company's ambition includes doubling its workforce across various departments, such as data centers and AI agents, as well as increasing private hiring of chip-design engineers to develop its own AI hardware, potentially reducing reliance on suppliers like Nvidia and Huawei.

Written by urgent.news from The Indian Express's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at indianexpress.com →

More in AI

AI Text Watermarking: The Statistics Hiding Inside Every Sentence

Something changed on August 2, 2026. Every new Claude model started embedding an invisible mark into the text it generates. You can't see it, strip it with backspace, or find it with Ctrl+F.

  • Claude models embed invisible statistical bias watermark since August 2026
  • Google's SynthID uses pseudorandom token sampling for watermarking
  • Watermark detection requires secret key and one-proportion z-test

More from Saturday 15 August →