DeepSeek resurfaces with cheap, capable V4

2026-04-27 · Source: The Rundown AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, medium

Summary

DeepSeek, a Chinese AI lab, has released preview versions of its V4 open-source AI models, featuring 1M-token context windows and support for Huawei Ascend chips. The V4 Pro model demonstrates strong performance, with DeepSeek's internal evaluations placing it near GPT-5.4 and Gemini 3.1-Pro in reasoning. External benchmarks show V4 Pro topping Vals AI's Vibe Code Bench, though it ranks in a fourth tier on AA’s Intelligence Index alongside Meta’s Muse Spark. Notably, V4 Pro offers significantly lower pricing at $1.74 per 1M input tokens and $3.48 per 1M output tokens, compared to GPT-5.5's $5/$30 and Opus 4.7's $5/$25. The integration with Huawei's Ascend chips provides a viable alternative to Nvidia's AI infrastructure, highlighting a strategic development in the global AI chip market.

Key takeaway

For CTOs and AI Engineers evaluating large language models, DeepSeek V4 presents a compelling option due to its competitive performance and significantly lower pricing. Its support for Huawei Ascend chips also offers a strategic alternative to Nvidia-dependent infrastructure, potentially mitigating supply chain risks and reducing operational costs. You should explore V4 for projects where budget efficiency and hardware diversification are critical considerations.

Key insights

DeepSeek V4 offers competitive AI model performance at lower costs, supported by Huawei chips.

Principles

• Cost-efficiency drives AI adoption.
• Diverse hardware support reduces dependency.

In practice

• Evaluate DeepSeek V4 for cost-sensitive AI projects.
• Consider Huawei Ascend chips for alternative AI infrastructure.

Topics

• DeepSeek V4
• AI Models
• Huawei Ascend Chips
• AI Pricing
• Anthropic Project Deal

Best for: CTO, VP of Engineering/Data, AI Engineer, Tech Journalist, Director of AI/ML, General Interest

Related on AIssential

• China’s DeepSeek Just Dropped a New AI Model Built on Huawei Chips — DeepSeek's V4 model demonstrates high-performance AI can be developed using domestic Chinese hardware, challenging U.S. chip dominance.
• DeepSeek V4 - almost on the frontier, a fraction of the price — DeepSeek V4 models offer competitive performance and industry-leading low costs through significant efficiency gains.
• [AINews] DeepSeek V4 Pro (1.6T-A49B) and Flash (284B-A13B), Base and Instruct — runnable on Huawei Ascend chips — DeepSeek V4 advances open-weight long-context and agentic coding through novel attention mechanisms and efficient architecture.
• How Deepseek v4 Connects to the US Grid — AI compute scarcity drives strategic equity-for-compute deals and necessitates national infrastructure investment.
• The Download: DeepSeek’s latest AI breakthrough, and the race to build world models — DeepSeek's V4 model and the rise of world models signify key advancements in AI capabilities and strategic independence.
• DeepSeek-V4 arrives with near state-of-the-art intelligence at 1/6th the cost of Opus 4.7, GPT-5.5 — DeepSeek-V4 offers frontier-class AI performance at dramatically reduced costs through architectural innovation and open-source licensing.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Rundown AI.