DeepSeek-V4.1-Flash
DeepSeek's new large model, free to download, that can read images and text and is built for developers and long, complex tasks.
trckr found this on Nov 18, 69 days after it came out, so it sent no alert. Its release date comes from Hugging Face.
Key facts
- Released
- Sep 10, 2026
- What it does
- Text and chat: Writes, answers questions and holds a conversation.Thinks first: Can reason step by step before it answers: slower, but better at maths, code and tricky questions.Sees images: Understands pictures, screenshots and charts you share with it.
View moreView less
- Knowledge cutoff
- May 2025 · the latest events it learned about
Can I use it?
In an app
No app mentioned so far.
For developers
Developers can use it now through an API: How developers plug a model into their own apps and pay per use. A model that's "in the API" can be used by programmers, even if it isn't in a chat app yet..
Download
Free to download and run on your own computer: it has Open weights: The model's files are public, so anyone can download it, run it on their own computer and build on it, usually for free. trckr labels these "Open"..
Price and limits
What developers pay to use it through the API: How developers plug a model into their own apps and pay per use. A model that's "in the API" can be used by programmers, even if it isn't in a chat app yet., per million Tokens: The small chunks of text a model reads and writes, about three-quarters of a word each. Prices and limits are counted in tokens. sent (input) or written back (output), and how much it can take in at once.
- Input
- $0.15
- per million tokensCached: $0.003
- Output
- $0.60
- per million tokens
- Context window: How much a model can keep in mind at once: your question, any files you share and its own replies. Measured in tokens; 1,000 tokens is about 750 words.
- 1.05M tokens
- about 790,000 words
View moreView less
- Longest answer
- 393K tokens (about 290,000 words)
From the lab's own listing on models.dev and OpenRouter, checked .
How it scores
From independent testers, never the lab's own claims.
- Combined test score#23 of 239
- GPT-6.2 Astra174
- 2Claude Fable 5.2174
- 22Claude Opus 4.8160
- 23DeepSeek-V4.1-Flash159
- 24Gemini 3.8 Flash159
Source: Epoch AI · Epoch Capabilities Index
- Chat answers, by people's votes#36 of 87
- Gemini 4 Carbon1,526
- 2Gemini 4 Argon1,521
- 35GPT-6 Astra1,475
- 36DeepSeek-V4.1-Flash1,475
- 37Grok 4.20 Beta 11,475
Source: LMArena · LMArena text, adjusted for style
- Building apps, by people's votes#20 of 87
- Claude Fable 5.21,824
- 2Claude Opus 5.51,818
- 19GPT-6 Sol1,674
- 20DeepSeek-V4.1-Flash1,659
- 21Kimi K31,658
Source: LMArena · LMArena WebDev
Compare DeepSeek-V4.1-Flash with
It came out 28 days after DeepSeek V4 Pro 0813, the lab's previous model.
The details
DeepSeek-V4.1-Flash is a 760 billion parameters: Data-centre scale: it runs on many chips (GPUs) at once. Parameters are the numbers a model learns: a rough measure of its size. Multimodal: Works with more than text: a multimodal model can also look at images, listen to audio or watch video. Mixture of experts (MoE): A design that splits a big model into many "expert" parts and uses only a few for each word. It gives big-model quality at closer to small-model speed. model that processes images and text and generates text, with a technical report describing architecture changes that reduce memory use for long contexts. It is released under the MIT license and available as a free download on Hugging Face: The website where labs and researchers publish models anyone can download. Think of it as an app store for AI models..
Model ID: The exact name developers type to use a model, like gpt-5-mini. New IDs often show up in a lab's systems before the official announcement, which is how trckr spots many releases early.
deepseek/deepseek-v4.1-flashdeepseek-ai/DeepSeek-V4.1-FlashTimeline
/GitHub huggingface/transformersFirst signal
Get alerts like this the moment they happen.
From DeepSeek and every other major lab.