:

DEEPSEEK LAUNCHES ULTRA-LOW-COST AI MODEL

AI DESK1 MIN READ
THU, SEP 10, 2026

■ AI-SUMMARIZED FROM 1 SOURCE ▸ TIMELINE

DeepSeek released a new AI model priced at a fraction of a cent per million tokens, intensifying competition against OpenAI, Anthropic, and other rivals in the large language model market.

The Chinese AI company's latest offering undercuts competitors significantly on pricing, marking another aggressive move in the increasingly crowded LLM space. DeepSeek has positioned itself as a cost-focused alternative to established players, leveraging different technical approaches to achieve lower operational expenses. The model's minimal pricing strategy puts fresh pressure on companies like OpenAI and Z.AI, which maintain higher token costs for their services. Industry analysts note this reflects broader market dynamics where cost efficiency has become a primary competitive lever alongside model performance and reliability. DeepSeek's aggressive pricing follows previous releases that similarly challenged incumbent pricing models. The company's approach suggests a deliberate strategy to capture market share through affordability, potentially forcing competitors to reconsider their pricing structures. The move underscores how AI model commoditization continues to accelerate, with multiple players vying for dominance through cost leadership rather than premium positioning alone.

■ SOURCES

Bloomberg Tech

■ SUMMARY WRITTEN BY AI FROM THE LINKS ABOVE

■ MORE FROM THE AI DESK

NASA and IBM have jointly developed an artificial intelligence model designed to support lunar exploration during the Artemis program. The collaboration aims to enhance mission planning and data analysis for future moon missions.

JUST NOWAI Desk

Meta has unveiled Muse, an AI agent operating through WhatsApp that handles shopping, email composition, travel booking, and price negotiation for users. The service integrates Stripe's Link for payments, positioning Meta ahead of OpenAI in direct commerce capabilities.

JUST NOWAI Desk

Deepseek released V4.1-Flash, a multimodal model with 552 billion parameters that reduces KV cache memory to 25% of its predecessor. The model matches performance of larger competitors while using only 16 billion active parameters per token.

JUST NOWAI Desk

Ant International, Visa, and Mastercard are collaborating on a unified standard for payments processed by AI agents. The initiative responds to McKinsey's forecast of $3 trillion to $5 trillion in AI-driven commerce by 2030.

1H AGOAI Desk

■ SUBSCRIBE TO THE DAILY BRIEF

ONE EMAIL, 5 STORIES, 06:00 UTC. UNSUBSCRIBE ANYTIME.