Chinese artificial intelligence startup DeepSeek has introduced its latest AI model, DeepSeek-V4.1-Flash-1017, promising faster task execution, improved benchmark results, and greater cost efficiency. The update reflects the company's continued efforts to strengthen its position in the competitive global artificial intelligence market.
Announced on Saturday, October 17, the upgraded DeepSeek AI model is now accessible through the company's official website, mobile application, and application programming interface (API). DeepSeek confirmed the release through posts on its official X account, highlighting improvements designed to enhance productivity and streamline AI-powered workflows.
The company described DeepSeek-V4.1-Flash-1017 as delivering a “Pareto improvement” in performance and cost, meaning users can benefit from stronger capabilities without facing higher operating expenses. The model also generates shorter responses and completes tasks more quickly, potentially reducing computational demands for developers and businesses.
DeepSeek said the updated model has been optimized for several AI development platforms, including DeepSeek Harness, Tencent CodeBuddy, and WorkBuddy. These integrations aim to improve automated programming, software development, and complex project management.
A major feature of the new DeepSeek model is its use of reinforcement learning with self-feedback. This technology allows the AI system to refine its performance during complex development processes, including creating video games and generating three-dimensional models from scratch.
Alongside the model upgrade, DeepSeek expanded access to its Search and Fetch APIs, making both services available to all users. Powered by the company's proprietary search engine, the APIs are priced at $0.75 per 1,000 calls for each endpoint, offering developers additional tools for integrating search capabilities into AI applications.
Benchmark results also indicate measurable improvements. DeepSeek-V4.1-Flash-1017 achieved a score of 77.0 on DeepSWE v1.1, compared with 74.2 for its predecessor. On Automation-Bench V1.0.6, the model scored 60.5, rising from the previous version's 54.8.
The latest DeepSeek AI update highlights the company's focus on delivering more capable, affordable artificial intelligence solutions. As competition intensifies across the AI industry, improvements in coding performance, automation, and API accessibility could help DeepSeek attract more developers and enterprise users worldwide.
