AI News: Everyone's Mad At Anthropic Now

2026-05-01 · Source: Matt Wolfe · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

DeepSeek V4, a new open-source model, was released with a 1 million token context length and near state-of-the-art performance, closely rivaling models like GPT 5.4 and Claude Opus 4.7 in benchmarks. Its pricing is significantly lower, at $1.74 per million input tokens and $3.48 per million output tokens, compared to GPT 5.5's $5 and $30 respectively. Other notable open models include Nvidia's Neotron 3 Nano Omni, a multimodal agent-focused model, and Poolside AI's Laguna XS2 (33B parameters) and M1 (225B parameters), with XS2 being open-weight. Mistral also released an open-weight 128B dense model for remote agents. Additionally, Alibaba introduced Quinn Image 2.0 Pro, and XAI launched Grok Voice ThinkFast 1.0, a low-latency voice model now powering Starlink's customer support.

Key takeaway

For CTOs and AI engineers evaluating large language models, the emergence of high-performing, open-source models like DeepSeek V4 presents a compelling alternative to proprietary solutions. You should assess these open models for potential cost savings and enhanced data privacy, especially for use cases where near state-of-the-art performance is sufficient. This shift could significantly reduce operational expenses and mitigate vendor lock-in.

Key insights

Open-source AI models are rapidly closing the performance gap with proprietary models while offering significant cost and privacy advantages.

Principles

• Open-weight models drive down AI service costs.
• Multimodal AI enhances agent capabilities.
• Ethical considerations impact AI adoption.

Method

Chinese AI companies are developing more cost-effective training methods due to export restrictions, enabling them to produce competitive open-weight models at lower prices than US-based frontier labs.

In practice

• Evaluate open models for cost-sensitive use cases.
• Consider local deployment for enhanced privacy.
• Test multimodal models for agent-based applications.

Topics

• Open-Source AI Models
• Large Language Models
• AI Ethics and Governance
• AI Pricing
• Multimodal AI

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

Related on AIssential

• What’s next for Chinese open-source AI — Chinese open-source AI models offer near-frontier capabilities at a fraction of the cost, rapidly reshaping global AI innovation.
• A lot has changed in 3 months..... — The increasing cost of commercial AI services is accelerating the adoption of efficient, local, and open-source models.
• Are Language Models a Commodity? — Language models are increasingly accessible and affordable, but advanced, specialized applications remain premium.
• The New Open-Weights Leader, Big AI’s Political Influence, Predicting Illness, Faster Reasoning — Open-weights AI models are rapidly closing the performance gap with proprietary frontier models, offering high-performance, customizable alternatives.
• Deepseek is a Problem — The US open-source AI model is failing due to lack of monetization, risking Chinese dominance and geopolitical influence.
• China’s DeepSeek previews new AI model a year after jolting US rivals — DeepSeek's V4 model aims to challenge top US AI systems, emphasizing coding prowess and domestic chip compatibility.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Matt Wolfe.