DeepSeek-V4.1-Flash-1017Open Weight
Released
DeepSeek-V4.1-Flash-1017 is an AI model released by DeepSeek on Saturday, Oct 17 2026, 37 days after DeepSeek-V4.1-Flash. It is an open-weight model — the trained weights are available to download and run. Benchmark results (shown below) cover NL2Repo-Bench, AutomationBench, DeepSWE 1.1, Terminal-Bench 4.0, Humanity's Last Exam, Agent's Last Exam, and 1 more.
Get DeepSeek releases and AI news.
Every other Monday
Benchmarks
Coding
NL2Repo-BenchRepo-level code generation — Tests whether the AI can turn a natural-language requirement into working code across an entire repository, not just produce a single function or patch. Higher is better.
68.2%
#1Best published NL2Repo score of all tracked models
DeepSWE 1.1Agentic coding — Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. (Version 1.1 of the test.) Higher is better.
77%
#4 of 36Best: Muse Wonder Full · 79%
Terminal & CLI
Terminal-Bench 4.0Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better.
42.3%
#11 of 29Best: Claude Sonnet 5.5 · 70.6%
Agentic & tool use
AutomationBenchBusiness workflows — Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better.
60.5%
#1Best published AutoBench score of all tracked models
Agent's Last ExamAgentic computer use — A hard set of desktop and operating-system tasks an AI agent has to finish by looking at the screen and working the machine itself. The score is the share it passes outright — partial credit does not count. Higher is better.
35.4%pass@1
#4 of 10Best: GPT-6 Astra · 59.3%
Reasoning & science
Humanity's Last ExamMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better.
42.6%no tools
#10 of 24Best: Claude Fable 5.1 · 60.9%
Multimodal
ChartographyChart tasks — A chart-centred test run with tools available to the AI, reported separately from the chart-reading benchmarks above it. Higher is better.
83%with tools
#4 of 6Best: Claude Opus 5.5 · 89%
Compare DeepSeek-V4.1-Flash-1017 with
Suggested comparisons
DeepSeek-V4.1-Flash-1017vsDeepSeek-V4.1-FlashDeepSeek-V4.1-Flash-1017vsGPT-6.1 SolDeepSeek-V4.1-Flash-1017vsClaude Haiku 5.5DeepSeek-V4.1-Flash-1017vsGemini 4 ArgonDeepSeek-V4.1-Flash-1017vsMuse Spark OpenDeepSeek-V4.1-Flash-1017vsGrok 4.7DeepSeek-V4.1-Flash-1017vsMistral Large 4DeepSeek-V4.1-Flash-1017vsKimi K3DeepSeek-V4.1-Flash-1017vsGLM-5.3-FlashDeepSeek-V4.1-Flash-1017vsQwen3.8-Max-0902DeepSeek-V4.1-Flash-1017vsNemotron 3.5 Super VLFrequently asked questions
DeepSeek-V4.1-Flash-1017 was released by DeepSeek on Saturday, Oct 17 2026.