# DeepSeek-V4.1-Flash DeepSeek-V4.1-Flash is an AI model released by DeepSeek on Sep 10 2026. It has a 1M context window and open weight. Tracked results include 30% on Terminal-Bench 3.0, 90.6% on Terminal-Bench 2.1 and 88.1% on CyberGym. ## Facts | Field | Value | | --- | --- | | Model | DeepSeek-V4.1-Flash | | Developer | DeepSeek | | Release date | Thursday, Sep 10 2026 | | Licensing | Open Weight | | Context window | 1M | ## Tracked benchmark scores | Benchmark | Score | Source | What it measures | | --- | --- | --- | --- | | BullshitBench v2 | 57% | [BullshitBench](https://github.com/petergpt/bullshit-benchmark) | Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | | DeepSWE 1.1 | 74.2% | Lab | 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. | | NL2Repo-Bench | 65.4% | Lab | 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. | | Terminal-Bench 4.0 | 31.2% | Lab | 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. | | Terminal-Bench 3.0 | 30% | Lab | command-line task completion (v3.0, much harder task set) | | Terminal-Bench 2.1 | 90.6% | Lab | Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | | CyberGym | 88.1% | Lab | Tests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better. | | Humanity's Last Exam (no tools) | 36.8% | Lab | 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. | | Humanity's Last Exam (with tools) | 63.9% | Lab | Humanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | | GPQA Diamond | 90.9% | Lab | Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | | Agent's Last Exam (pass@1) | 31.8% | Lab | 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. | | AutomationBench | 54.8% | Lab | 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. | | Chartography (with tools) | 78.9% | Lab | A chart-centred test run with tools available to the AI, reported separately from the chart-reading benchmarks above it. Higher is better. | | threejseval | 1642 | [threejseval](https://threejseval.com/ranking) | Every model gets the same prompt — "the Eiffel Tower", "a glass fishbowl", "a robot arm picking toys into a box" — and builds a 3D scene in Three.js. Real people then see two scenes side by side, names hidden, and vote for the one they prefer. The votes become a chess-style Elo rating on threejseval.com, averaged across all the prompts. It measures whether the scene looks and moves right to a human eye, not whether the code passes a test. Higher is better. | ## About DeepSeek-V4.1-Flash DeepSeek V4.1 Flash was introduced in September 2026 with a new causal encoder-decoder architecture. DeepSeek reported 552 billion total parameters, with about eight billion active for input encoding and 16 billion active for output generation. The model was designed to improve inference speed and throughput while supporting native visual understanding. The release brought this architecture to DeepSeek’s hosted assistant and API, with downloadable weights for self-managed deployment. Its vision capability belonged to the model rather than to a text-only endpoint supplemented solely by application tools. The 4.1 name also distinguishes it from the earlier V4 Flash checkpoints, including a separate experimental vision release in the V4 line. ## Questions and answers ### When was DeepSeek-V4.1-Flash released? DeepSeek-V4.1-Flash was released by DeepSeek on Thursday, Sep 10 2026. ### Who made DeepSeek-V4.1-Flash? DeepSeek-V4.1-Flash was built by DeepSeek. Chinese AI lab known for efficient, open-weight models. Gained attention for strong performance at lower cost. ### What benchmark scores did DeepSeek-V4.1-Flash get? DeepSeek-V4.1-Flash reports 14 tracked benchmark scores — BullshitBench v2: 57%; DeepSWE 1.1: 74.2%; NL2Repo-Bench: 65.4%; Terminal-Bench 4.0: 31.2%; Terminal-Bench 3.0: 30%; Terminal-Bench 2.1: 90.6%; CyberGym: 88.1%; Humanity's Last Exam (no tools): 36.8%; Humanity's Last Exam (with tools): 63.9%; GPQA Diamond: 90.9%; Agent's Last Exam (pass@1): 31.8%; AutomationBench: 54.8%; Chartography (with tools): 78.9%; threejseval: 1642. Tracked scores may come from lab reports or independent benchmarks; source details accompany the benchmark data. It holds the best score among all models tracked here on Terminal-Bench 3.0, Terminal-Bench 2.1 and CyberGym. ### What is the context window of DeepSeek-V4.1-Flash? DeepSeek-V4.1-Flash has a context window of 1M. Providers may specify separate input and output token limits. ### Is DeepSeek-V4.1-Flash open source? Partly. DeepSeek-V4.1-Flash is an open-weight model: the trained weights are free to download and run locally or on your own infrastructure, but the training data and code are not fully released and the license may restrict some commercial uses. It is not open source in the strict sense. ### What came before and after DeepSeek-V4.1-Flash? DeepSeek's previous tracked release was DeepSeek-V4-Pro-0813 on Aug 13 2026, 28 days earlier. It was followed by DeepSeek-V4.1-Flash-1017 on Oct 17 2026. --- Canonical page: https://aireleasetracker.com/model/deepseek/deepseek-v4.1-flash Site index: https://aireleasetracker.com/llms.txt Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.