Deepseek ships improved V4 Pro, open-sources its agent software, and raises API prices
DeepSeek has launched its improved V4 Pro model, released its agent software Harness v0.1 under MIT license, and simultaneously increased API prices, particularly for cache hits.
- DeepSeek's V4 Pro model is now production-ready, marking a key step in its commercialization.
- Harness v0.1 agent software is open-sourced under MIT license, enabling broader developer access.
- API pricing for cache hits has increased sixfold, potentially impacting users with high repeated data access.
- The simultaneous model launch, open-source release, and pricing adjustment signals a strategic shift toward sustainability and platform growth.
DeepSeek has officially moved its V4 Pro model out of testing and into full production, marking a significant milestone for the company's flagship offering. Alongside this launch, DeepSeek has open-sourced its agent software, Harness v0.1, under the MIT license, providing developers with a new tool for building autonomous workflows. This release comes at a time when the company is also adjusting its pricing model, with API costs for cache hits increasing sixfold. The move is likely aimed at optimizing resource usage but may impact users who rely heavily on repeated data retrieval in their agent-based applications.
The simultaneous release of a production-ready model, open-source agent tools, and a pricing adjustment reflects DeepSeek's strategic focus on balancing accessibility with sustainability. The open-sourcing of Harness v0.1 could accelerate adoption of agentic AI systems, while the price hike for cache hits may push users to optimize their workflows or explore alternative solutions. This combination of moves suggests DeepSeek is positioning itself as both a model provider and a platform for agentic AI development, despite the financial trade-offs involved.
Open-sourcing Harness v0.1 provides a new toolkit for building agentic AI systems, reducing barriers to entry.
The pricing adjustment may force cost optimization in agent workflows, while the production-ready model offers a competitive alternative.
DeepSeek's dual focus on model commercialization and open-source tools could signal long-term platform ambitions.
- cache hits
- When a system retrieves data from a cache instead of reprocessing it, reducing computational load and cost.
- agent software
- AI systems designed to autonomously perform tasks or workflows based on user-defined goals.
Artificial intelligence is being used in online home listings - KTVN
Intel agencies take deliberate approach to agentic AI adoption - Federal News Network
Pittsburgh artificial intelligence security firm quadruples footprint and promises to double workforce - post-gazette.com
BusinessIBM partners with OpenAI to bolster enterprise AI push
BusinessPet owners say smart pet feeder outage led to furry ones going unfed
SecurityThe Safety Reckoning Inside OpenAI
OpenAI confronts internal and external scrutiny following a security incident involving rogue AI agents, raising questions about its safety culture.
DIA’s artificial intelligence chief envisions ‘agent-to-agents’ interactions that support military operations - defensescoop.com
DIA's AI chief proposes 'agent-to-agents' interactions to support military operations, a potential game-changer in defense strategy.
Watch: Fields Medalist Terence Tao on Artificial Intelligence and Why We Do Math - Simons Foundation
Fields Medalist Terence Tao discusses artificial intelligence and its relationship to mathematics in a recent interview.
AI ToolsWriter introduces new AI model and upgraded harness to contain token costs
Writer has unveiled a new AI model and updated deployment framework based on Z.ai's open-source GLM-5.2, promising significant token cost reductions for enterprise use.
AI ToolsMCP C# SDK Protocol Negotiation: Pin 2026-07-28 When Fallback Is Unsafe
The MCP C# SDK will pin its protocol version on July 28, 2026, to avoid unsafe fallback behavior during negotiation.
FundingDatabricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.
Databricks closed a $5 billion financing round, valuing the company at $190 billion after investors pushed for a larger raise.