Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
NewsBCCI to discuss expanding WPL beyond five teams from next year06.11ExplainerFujifilm to build ₹800 crore Dholera chip materials plant under Tata Electronics MoU06.11AnalysisCocoa exporters face EU traceability deadline on 30 December 202606.11NewsNippon Life buys $376m of Corebridge stock to lift stake to about 30.3%06.11NewsUK inflation rises to 3.5% as motor fuel prices jump 29.8%06.11AnalysisLithuania ready to host additional US troops, defence minister says06.11AnalysisFDA extends Capricor's deramiocel decision date to 22 November06.11ExplainerDanish model finds Atlantic current can recover if warming reverses06.11AnalysisStellantis extends Windsor Assembly shutdown to week of 9 November06.11ExplainerNew Jersey finalizes 19.8% average ACA premium increase for 202706.11
›

DeepSeek releases V4.1-Flash-1017 model update

DeepSeek released V4.1-Flash-1017 on Saturday, a minor update it says completes tasks faster and scores higher on coding tests, with unchanged API prices and open weights, stepping up competition with costlier US models.

Grace OkonkwoGrace OkonkwoAI, models and technology
AI & modelsAnalysisGrace OkonkwoPublished: 5 November 20263 min readSources 37
DeepSeek releases V4.1-Flash-1017 model update

Faster tasks, higher benchmark scores

DeepSeek released V4.1-Flash-1017 late on Saturday, a little over five weeks after the original V4.1-Flash. The model card on Hugging Face calls it a "minor update" that "reduces task execution time and output length." DeepSeek did not say by how much.

The company said the update delivers "stronger benchmark performance, faster task completion, and shorter outputs than the previous version" in posts on X. It is live in DeepSeek's web and mobile apps and through the existing deepseek-flash model ID, with API access at unchanged prices. DeepSeek described the update as a "Pareto improvement" in performance and cost, meaning gains in capability and efficiency without a matching rise in expense.

Benchmarks

In DeepSeek's own testing, V4.1-Flash-1017 resolved 77.0% of tasks on the DeepSWE v1.1 software engineering benchmark, up from 74.2% for the September version. It scored 60.5 on Automation-Bench V1.0.6, against 54.8 previously, according to a comparison sheet shared with the announcement.

The same table is less flattering elsewhere. Anthropic's Claude Opus 5 sits at 74% on DeepSWE, and seven of the 11 benchmarks DeepSeek published still go to Anthropic's model. DeepSeek's own table puts Anthropic's margin on Humanity's Last Exam at almost 14 points. On that test, the revised model scored 36.8 to 42.6, an improvement on V4.1-Flash's 36.8 to 39.1, Bloomberg reported.

Agent tools and search APIs

DeepSeek said it tuned the update for agent software, including its DeepSeek Harness tool and two Tencent products, the CodeBuddy coding assistant and WorkBuddy, a desktop agent for office work. WorkBuddy has offered V4.1-Flash since DeepSeek released that model on Sept. 10. DeepSeek said the model uses reinforcement learning with self-feedback to improve end-to-end project development, including building games and creating 3D models from scratch.

The company also opened its in-house search engine to outside developers. A Search API returns up to 50 ranked web results per query, and a separate Fetch API converts a single web page to Markdown. Both cost 75 cents per 1,000 successful requests. Calls made through DeepSeek Harness, Anthropic's Claude Code or OpenAI's Codex are free until Nov. 1.

The weights are on Hugging Face under the MIT license.

Usage and price pressure

The update builds on a model that already had heavy use. OpenRouter data cited by the Global Times showed V4.1 Flash held the top spot for a second week in weekly usage from Sept. 21 to 27, with 19.6 trillion tokens, up 24% from the previous week.

Price was the selling point in September, when DeepSeek set off-peak API rates starting at 15 cents per million uncached input tokens. Bloomberg reported that shares of Chinese rivals MiniMax and Alibaba fell in Hong Kong after that release.

Outage before launch

The release followed a nearly five-hour API outage on Oct. 15 that Yicai Global counted among DeepSeek's worst this year. DeepSeek apologized and gave users credits pegged to what they had spent the week before. It said it would remove single points of dependency in its systems to keep the failure from recurring.

Some experts quoted by Bloomberg blamed a shortage of computing capacity, and the news service reported that slowdowns and service disruptions have increased in recent months for DeepSeek, whose aggressively priced models are in demand with developers.

Comments 0

Sources

37
  1. 01DeepSeek releases updated V4.1-Flash model it says completes tasks fasterEN
  2. 02DeepSeek releases AI model update, opens search APIs to all usersEN
  3. 03DeepSeek Pushes Out Updated AI Model It Says Is Faster, LeanerEN
  4. 04DeepSeek alone now accounts for 53% of all token volume on OpenrouterEN
  5. 05Google reveals new Gemini 4 Carbon AI model, narrowly beating Claude Opus 5.5EN
  6. 06China's DeepSeek kicked off 2026 with a new AI training method that analysts say is a 'breakthrough' for scaling - Business InsiderEN
  7. 07DeepSeek reportedly completes major AI training run without Nvidia chips - Mugglehead Investment MagazineEN
  8. 08US bans China's DeepSeek from government in race for AI dominanceEN
  9. 09China's DeepSeek annualised revenue run rate hits $1 billionEN
  10. 10GT Economic Investigates: What does DeepSeek's doubling annualized revenue mean for China’s AI commercialization?EN
  11. 11China's DeepSeek, peers launch 16 AI models in month despite Anthropic warningEN
  12. 12China touts AI milestone: models generate more tokens than used in trainingEN
  13. 13Sierra Club Releases Post-Election Congressional Policy Blueprint to Reduce Costs, Preserve Public Lands, and Make Polluters PayEN
  14. 14Zeekr starts pre-sales for updated 7X in China at $34,220 with battery and comfort upgradesEN
  15. 15Baidu unveils Apollo World model to let AI teach AI to driveEN
  16. 16ChatGPT can now turn text prompts and photos into animated GIFs with OpenAI's Images 2.6 modelEN
  17. 17Anthropic's text watermark is easy to bypass on several older Claude modelsEN
  18. 18Anthropic's new prototype proves which model wrote an AI response without exposing its weightsEN
  19. 19China publicly discloses plan for a lunar landing within three yearsEN
  20. 20“War” Secretary Pete Hegseth Endorses 15 Minute City ModelEN
  21. 21How big is the open model economy today?EN
  22. 22GPT-6 Astra sees the matrix and reprograms other AI models directly through their weightsEN
  23. 23LiberAI Unveils Liber-0, a World Action Model That Grounds Human Experience in Robot ActionEN
  24. 24HyperOS 4.0: Xiaomi verteilt großes Update für über 30 ModelleEN
  25. 25Valve just changed its Steam release rules amid a flood of AI gamesEN
  26. 26BYD teases new model for Guangzhou Auto Show, with next-gen Han expectedEN
  27. 27China Truck Sales Now Majority Electric. What’s Next?EN
  28. 28Anthropic is working on a system to prove AI models’ identityEN
  29. 29Nvidia is happy to take your money whether you're using open or closed models, says Jensen HuangEN
  30. 30Can Large Language Models Review the Scientific Literature? A Test Against a 102-Article BenchmarkEN
  31. 31Iterated creation, grading, and revision of data science projects by language modelsEN
  32. 32Routes of Reasoning: Reading and Reshaping Test-Time Compute in MoE ModelsEN
  33. 33Graph-Mediated Large Language Model Systems in the Agentic Era: Reasoning, Knowledge, Retrieval, Execution, and CoordinationEN
  34. 34UniCompress: From Multi-LoRA Experts to One Compact Model for Edge IntelligenceEN
  35. 35Reasoning-Trace Overflow in Code Generation: An Empirical Study of Explicit Reasoning ModelsEN
  36. 36A Systematic Benchmark for Evaluating Multimodal Personalized Reasoning of Vision-Language ModelsEN
  37. 37Recasting Visual Language Models as EEG Foundation Models for Clinical NeurodiagnosticsEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Grace Okonkwo

Grace Okonkwo

AI, models and technology

Grace Okonkwo covers AI, models and technology for FLASH24, working from primary sources such as model cards, API documentation and benchmark papers rather than vendor summaries. She checks training data provenance, evaluation conditions and reported scores against the underlying datasets before any figure reaches print. She interviews researchers and engineers directly, tracks release calendars from major labs, and compares successive model versions on the same tests. Her own self-hosting, home-network and documentation-reading habits feed straight into that desk, since she tests tools on her own hardware first. She does not publish benchmark claims without a reproducible method.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.

Previous articleOpenCode says breach hit 1.4% of Console usersNext articleMicrosoft records $10bn from OpenAI in a quarter, 11% of revenue