DeepSeek V4.1 Flash
A lightweight, fast-inference model in the DeepSeek series belonging to the Flash family. It features lower inference costs and is suited for high-frequency invocation scenarios.
Core Facts
The DeepSeek Flash version is a distilled model optimized for low-latency scenarios, trading some complex reasoning capability for faster response speeds and lower token unit prices.
DeepSeek V4 Flash is an MoE multimodal model with a total parameter count of 284B.
DeepSeek V4 Flash is a cost-effective inference model released by DeepSeek, optimized for speed and cost.
Timeline (last 90 days)
Particle-1 is post-trained on DeepSeek V4.1 Flash.
In the Intelligence Index V5 released by Artificial Analysis, Cloud Opus 5.5 ranks first with 58 points, and DeepSeek V4.1 Flash ties for first among open-weight models with 32 points.
According to Artificial Analysis data, DeepSeek V4.1 Flash can reach 1286 tokens per second via Lithos AI, and Base 10 provides an inference speed of 845 tokens per second.
In a local coding test, DeepSeek V4.1 Flash, from a minimal prompt, generated a 3D Rubik's Cube simulation with correct colors and accurate rotation physics, and both scrambling and solving completed normally.
In the pastry counting task, the model proactively inferred occluded pastries, giving a reasonable total range of about 130 to 145; when asked a follow-up, it added about 21 occluded macarons for a total of 45, with the follow-up taking 23 seconds.
In a local coding test, DeepSeek V4.1 Flash, with only the single prompt "build an Excel clone for the browser," worked for 50 minutes on the DJX workstation and produced a spreadsheet application called Gridline that supports double-click editing, formula calculation, bolding, drag-to-fill, chart insertion, and drag-and-drop operations.
DeepSeek V4.1 Flash scored 42 on Humanities Last Exam, tied with Kimi K3 and GLM 5.3; its Terminal Bench score was also 42.
In the blogger's test, the interactive 3D mechanical watch webpage generated by DeepSeek V4.1 Flash, named Mechanica, had accurate part positions, and hovering over the hairspring displayed an explanation, but the watch back lacked the oscillating weight for automatic winding.
In the blogger's test, DeepSeek V4.1 Flash's output for a 3D model of a foldable phone was better than the same prompt output from Kimi's new model, but the folding animation had a minor issue where it briefly moved out of the frame after clicking.
In a browser tool-use test, DeepSeek V4.1 Flash analyzed a Sam Altman interview video and, after working for 6 minutes and 24 seconds, extracted five clips suitable for editing and gave timestamps verified as accurate by the blogger.
40 more timeline events
All Facts (20)
The DeepSeek Flash version is a distilled model optimized for low-latency scenarios, trading some complex reasoning capability for faster response speeds and lower token unit prices.
90%VerifiedDeepSeek V4 Flash is an MoE multimodal model with a total parameter count of 284B.
80%VerifiedDeepSeek V4 Flash is a cost-effective inference model released by DeepSeek, optimized for speed and cost.
80%VerifiedDeepSeek V4.1 Flash claims to achieve breakthroughs across three dimensions: capability, speed, and cost.
75%VerifiedV4.1 Flash is natively multimodal in design, capable of directly viewing and understanding image content.
75%VerifiedDeepSeek is deploying native multimodal capabilities at the Flash level rather than the flagship version first, with the intent to prioritize capturing high-frequency, large-scale application scenarios.
75%VerifiedDeepSeek V4.1 Flash scores lower than the previous version on GPQA, HLE, and other benchmarks.
75%VerifiedCompared to the previous-generation V4 Flash, V4.1 Flash reduces the demand for high-bandwidth memory (HBM) to one-quarter and the demand for SSD persistent storage to one-eighth.
75%VerifiedV4 Flash's cache size is only 7% of that of V3.2.
70%VerifiedV4.1 Flash adopts a new CED (Causal Encoding) architecture, whose core idea is an asymmetry in computation between input and output.
70%VerifiedV4.1 Flash surpassed Opus 5 on the DeepSWE benchmark by a narrow margin of 0.2 points.
70%VerifiedThe Flash version is positioned as a lighter, faster inference option suited for latency-sensitive or resource-constrained scenarios, while the PRO version targets a higher capability ceiling with greater demands on VRAM and compute
70%VerifiedArtificial Analysis evaluation shows DeepSeek V4.1 Flash scored 40 on the Intelligence Index, surpassing V4 Pro to become DeepSeek's new flagship, featuring native multimodal input, 8B activated parameters on input and 16B on output, released under the MIT license.
65%VerifiedV4.1 Flash activates 8 billion parameters when processing input and activates 16 billion parameters when generating output.
65%VerifiedCompared to the previous generation Flash, V4.1 Flash's high-bandwidth memory requirements have been reduced to 1/4, and solid-state storage requirements have been reduced to 1/8.
65%VerifiedThe browser operating system interface implemented by DeepSeek V4.1 Flash presents a dynamic desktop with animation effects, features refined popup designs, and Dock bar applications can be invoked normally.
65%VerifiedDeepSeek V4.1 Flash adopts a Causal Encoder-Decoder asymmetric architecture with a total parameter count of 552B.
65%VerifiedCompared to the old Flash model, V4.1 Flash has more comprehensive capabilities and more accurate recognition, but takes longer and costs more per session.
65%VerifiedInterfaces that previously called DeepSeek V4 Pro will automatically switch to the new V4.1 Flash model, requiring no code changes from developers.
65%VerifiedV4.1 Flash activates only 8 billion parameters during the input stage and 16 billion parameters during the output stage.
65%