Journalism begins where hype ends

,,

The use of AI is never a purely technical matter: when it enters processes that affect people’s lives, it touches on rights, opportunities, status and freedom"

—Pope Leo XIV

DeepSeek-V4.1- Flash-1017 Tops Arena AI’s Code Arena WebDev Leaderboard, Beats Kimi K3 at Lower Cost

DeepSeek-V4.1-Flash-20261017 has taken the top open-model spot on Arena AI’s Code Arena WebDev leaderboard, scoring 1,670 points and ranking 12th overall. The model overtakes Kimi K3 (Max) while costing roughly one-tenth as much to run, strengthening DeepSeek’s position in cost-efficient AI coding.
DeepSeek AI whale logo for the DeepSeek-V4.1-Flash-1017 model launch.
October 19, 2026 09:00 AM IST | Written by Supriya Singh | Edited by Vaibhav Jha

Just two days after its launch, DeepSeek-V4.1- Flash-1017 has become the top open model in Code Arena WebDev leaderboard, surpassing Kimi K3 model, owing to the “Pareto improvement” in model’s agentic front-end tasks performance and lower cost.

According to Arena AI, DeepSeek Flash v4, which was updated by Chinese startup DeepSeek AI on October 17, has quickly risen in ranks among top open models scoring 1670 points and ranking 12th overall.

 

The AI model edges past Kimi K3 (Max) while running at roughly one-tenth the cost, priced at $0.30 per million input tokens and $1.20 per million output tokens.

On October 17, DeepSeek launched DeepSeek-V4.1- Flash-1017, an updated version of its AI model that the company says delivers stronger benchmark performance, faster task completion and shorter outputs than its predecessor.

Built as a Mixture-of-Experts architecture , the new model is available on DeepSeek’s website, mobile app and application programming interface. The company has described the update as a “Pareto improvement” in performance and cost, indicating that it aims to improve performance while maintaining or reducing the cost of using the model.

 

DeepSeek-V4.1- Flash-1017 is designed to support end-to-end project development, including building complete projects from scratch. According to the company the model uses reinforcement learning with self-feedback to improve its ability to execute complex tasks, ranging from game development to 3D modelling.

Architecturally, it builds on the V4.1-Flash multimodal Mixture-of-Experts design (approximately 552B backbone parameters plus Engram components). It uses a Causal Encoder-Decoder structure that activates only about 8B parameters during prefill and 16B during decode, supporting a 1-million-token context window and native image-text inputs.

The model has also been optimized for AI development tools, including DeepSeek Harness and Tencent’s CodeBuddy and WorkBuddy. DeepSeek has recommended upgrading to DeepSeek Harness version 0.2.1-rc.1 or later to get the best experience with the new model.

Alongside the model launch, DeepSeek has opened its Search and Fetch APIs to all users. The APIs are powered by the company’s in-house search engine, which says it handles hundreds of millions of requests worldwide everyday.

The company has priced each API endpoint at $0.75 per 1,000 calls. The newly available services are intended to enable developers to integrate search and information retrieval capabilities into their applications.

The launch expands DeepSeek’s offerings for developers seeking AI tools for coding, project creation and information retrieval.

Also Read: DeepSeek V4 Model Unveiled: China’s Cost Effective GenAI Push

Authors

  • AI FrontPage Reporter Supriya Singh

    Supriya Singh is a Reporter at AI FrontPage covering the AI & Education and AI & Jobs beats. She brings six years of print and digital experience, including three years at The Asian Age, where she reported on higher education, Delhi government, and crime. She is based in Delhi-NCR.

    LinkedIn
  • Vaibhav Jha, editor and co-founder at AI FrontPage

    Vaibhav Jha is an Editor and Co-founder of AI FrontPage. In his decade long career in journalism, Vaibhav has reported for publications including The Indian Express, Hindustan Times, and The New York Times, covering the intersection of technology, policy, and society. Outside work, he’s usually trying to persuade people to watch Anurag Kashyap films.

    LinkedIn

Related Posts

LATEST NEWS