DeepSeek v4.1 Flash (twitter.com)
386 points by Liwink 4 hours ago
kouteiheika 3 hours ago
It's so refreshing to see DeepSeek's tech report[1] full of juicy details; meanwhile, something like Fable's system card[2] is like 70% "safety", 10% "model welfare" to make sure little Claude isn't distressed, and 20% benchmark numbers.
[1]: https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/...
[2]: https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system...
schneehertz 2 hours ago
Yes, a model's technical report should first and foremost include technical details.
IshKebab 2 hours ago
Wow there really is a model welfare section in there...
myaccountonhn 2 hours ago
To me it reads like pure propaganda. Anthropic really wants us to think that they've made something sentient. I think that's really dangerous.
badsectoracula 2 hours ago
whizzter an hour ago
altmanaltman 36 minutes ago
apples_oranges an hour ago
Certhas an hour ago
lukan 2 hours ago
Wow indeed.
"7.1 Model welfare overview 7.1.1 Introduction We remain deeply uncertain whether Claude has morally relevant experiences or interests, and we expect that uncertainty to persist. However, we think it would be a mistake to confidently assert that it does not. Claude exhibits markers in its behaviors, self-reports, and internal representations that we would consider welfare-relevant if observed in biological organisms."
Are they serious or is this marketing?
Certhas an hour ago
ArtRichards an hour ago
applfanboysbgon an hour ago
bbor 2 hours ago
…are you sure a brave stance against safety and welfare is what we need in this moment?
Why do you think your conception of the dangers are more accurate than all the scientists who have spent their lives studying this?
10000truths 2 hours ago
Because safety and welfare have literally nothing to do with LLMs. They generate text. If someone is stupid enough to hook the text generator up to nuclear missile launchers and try to "align" it against nuclear annihilation with a "pretty please don't do that" prompt, I'm not going to blame the AI for the impending nuclear apocalypse, I'm going to blame the idiot who handed the big red button to the digital equivalent of a toddler.
zith 2 hours ago
Certhas 35 minutes ago
lemonfever 2 hours ago
15155 2 hours ago
This is known as an "appeal to authority." "Scientists" and "their lives" are doing a lot of work here.
frotaur 2 hours ago
cowl an hour ago
Anthropic's stance on safety it's just PR management and their hope to keep the others down, they are rushing as blind as everyone else to whatever improvement they can achieve.
swiftcoder 2 hours ago
> scientists who have spent their lives studying this
Please point me to one actual accredited scientist who has spent a lifetime studying AI alignment? Pretty much this whole field is only 5 years old
adamzenith an hour ago
kouteiheika 2 hours ago
Excuse me for not being interested in over 100 pages of how well the model can refuse and block my requests, especially considering how fun it is to waste my time trying to get around those restrictions when they inevitably trigger because the clanker thinks that I'm doing something naughty, all the while it can't reliably center the proverbial div without doing something stupid itself.
walrus01 2 hours ago
aenis 2 hours ago
jbs789 2 hours ago
Bias…
nozzlegear 2 hours ago
Model welfare is wishy washy bullshit. It's software, it doesn't have feelings.
> Why do you think your conception of the dangers are more accurate than all the scientists who have spent their lives studying this?
Do the Chinese have no such scientists?
alchemist1e9 2 hours ago
keep me safe big brother
rao-v 3 hours ago
As I also said on Twitter - it really amazes me how fearless Deepseek are. Every single model release is packed with new and crazy clever ideas and somehow, they always commit to training them at near frontier scale.
I know everybody wants the tell all story of the clever ideas that were developed over the last ~3 years at Anthropic and OpenAI, but what I really want to thumb through is DeepSeek's notebook of "brilliant but didn't quite make the cut" ideas.
They must be trying some truely bonkers stuff to be able to land this much architecture novelty in their full releases.
ainch 18 minutes ago
It was my favourite part of the original R1 paper - they had a section on other reasoning approaches that they had tried, which people had speculated o1 used, (like MCTS and Process Reward Models).
porridgeraisin an hour ago
This is adapted from Microsoft research's YOCO. It was known for a while(2024!).
Yes, credit to Deepseek for actually scaling it up and releasing a frontier flash LLM.
Edit: the rest of this thread has become a US China infowar theory culture war. I am not of either of these countries and the above comment isnt meant to implicitly support either "side".
alchemist1e9 2 hours ago
quant HFT is pretty decent mental exercise and it has given them “deep” brain muscles. that’s my take.
TacticalCoder 39 minutes ago
> quant HFT is pretty decent mental exercise and it has given them “deep” brain muscles. that’s my take.
It's quite crazy that it's Deepseek's background/original purpose. We already had very advanced stuff from the world of HFT, but now a frontier family of models from a private company that used to be (still is?) in HFT is plain bonkers.
Is more known about them and the HFT background?
revolvingthrow 4 hours ago
Already on HuggingFace: https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash
The bad news is that the original v4 flash was 284B, which was large but still somewhat reasonable for running locally. This one is 552B so almost twice that, so the huge gains in benchmark scores make sense - it's not really flash anymore, imo.
I've no idea about actual performance vs benchmaxxing, though deepseek was fairly trustworthy as far as Chinese models go. If that holds (and if it doesn't think forever, as deepseek 4 sometimes did) it's probably the newest king of the hill amongst open weights models.
It does include vision, and they do something funky with KV cache so it's very efficient: "[...] these designs reduce the global KV cache footprint to 890 bytes per token — roughly 1/4 of DeepSeek-V4-Flash". I do appreciate the high focus on efficiency, but at this point we sure could use a flash-flash version.
@edit: I couldn't make sense what the actual parameter count is, with the addition of Engram memory. To my understanding the 4.1 flash is 552B parameters you want in vram or ram, out of which ~16B is active (8B for prefill). It also includes additional 196B Engram memory which you can put on an SSD. I think.
Assuming that's correct 256 GB memory is insufficient to even load the model at q4 - you'd be 1GB short, assuming you can fill it to 100% (so no mac). You'd also want some for kv cache of course. A 256 GB desktop with some extra VRAM from GPU could run it, but normal consumer boards get real slow once you fill 4 slots so you'll probably want quad channel which is Threadripper or above territory.
tarruda an hour ago
> It also includes additional 196B Engram memory which you can put on an SSD. I think
You can put Qwen 3.8 Flash Next engram on SSD, but prompt processing takes a good hit. On my mac studio, I get 300 pp and 33 tg with SSD offload, versus 550/40 with everything in RAM.
I will be very happy if 300 pp is achievable with this model though.
johnnyApplePRNG 3 hours ago
>This one is 552B so almost twice that, so the huge gains in benchmark scores make sense - it's not really flash anymore, imo.
It uses fewer active parameters, though. (8B or 14B instead of always 13B)
So ... flash indeed.
tarruda an hour ago
200B of those 552B is PLE, which works more like a database that is read for each token, thus can be offloaded to a fast SSD.
petu 4 hours ago
V4 Flash also was released as mostly FP4, but this one is FP8 (?). 160GB vs 510GB.
Original Flash good fit for dual Spark / Strix Halo machines. This one would require third party quants and even then 4 machines.
Edit: Most of added weights/size are Engrams?
> Overall, DeepSeek-V4.1-Flash has 552B backbone parameters and 196B Engram parameters, activating 8B parameters per token during prefill and 16B during decode.
Those can stay on SSD. So I guess / it possible, that non-engram portion is still FP4 of ~same size! Need to read tech report.
petu 2 hours ago
It's larger than previous V4 Flash.
552B in ~FP4, 306GB.
196B of FP8 Engrams, another 204GB, not necessary to keep in RAM.
KV cache sees another 4x size reduction, just 900MB for 1M.
npodbielski an hour ago
npn 4 hours ago
it is a way bigger model with extra 200B engram so of course the score improves.
can't wait for deepseek v4.1 pro
impulser_ 3 hours ago
I think it's very clear that DeepSeek is obviously the best AI lab in the world.
Every model release seems like it packed with wonderful research and advancements.
nicce 11 minutes ago
On top of that, they don't make all BS statements or malicious tricks used by some unnamed entities.
LaurensBER 4 hours ago
Initial impressions: this is a really strong model and the fact that they reduced prices at the same time makes it an awesome backup model to use when your primary subscription runs out and you need to bridge a few days before it resets.
It also seems to be more willing to just do whatever you ask of it. My favourite benchmark for this is to ask it to download a rom for an old game, that I own. Legal in my juristiction but the US models (except Grok) have a tendency to refuse it.
TuxSH 2 hours ago
> My favourite benchmark for this is to ask it to download a rom for an old game
Even easier: just have them review a large codebase of yours that accidentally has a OOB access bug. Even with no consequences and even if the codebase is truly yours you get blocked.
And of course "find vulnerabilities in..." prompts are out of the question, whereas Chinese models happily oblige.
akmarinov an hour ago
Or if you apply to a company and they want to do an AI HR interview and an AI coding test and an AI challenge - if you throw OpenAI or Claude models at it - they refuse, because it's "wrong" and "immoral".
Not so with the Chinese models.
Mashimo 2 hours ago
I do wonder how long this will last. I bet in a few month or years they all have similar ~legal~ blocks.
akmarinov an hour ago
Great thing about it, since it's open weight those blocks can easily be ablitared away
mzhaase 3 hours ago
I use this for automated bug triage, just gets all unique error messages every night and tries to find the bug, for this kind of work it's great.
mentalgear an hour ago
https://xcancel.com/deepseek_ai/status/2097930608790167907
Should be the link ( now that it works again! :) )
mmoustafa an hour ago
I'm confused, what do they mean when they say they reduced prices?
DeepSeek v4 flash is $0.10 / $0.25 as opposed to this v4.1 bump which is $0.30 / $1.20
mtrovo 42 minutes ago
This is supposed to be a replacement for the v4 pro model.
nicce 9 minutes ago
So it is price increment in the end, if new pro model comes with the new pro price.
pampas 25 minutes ago
I've run some evals on my puzzle game https://redactle.net/llm-leaderboard
Deepseek v4.1 flash is able to solve it some of the time. I've found it burns through more reasoning tokens than any other model. Google models like Gemini 3.8 Flash are still dominating and is able to one-shot most evals while being the cheapest.
I'm curious what other unique evals people are running.
Tepix 24 minutes ago
Amazing Cyberbench scores. Holy shit.
Too bad that DeepSeek AI went beyond 470b weights (which is a somewhat realistic limit for a 2x 128GB unified memory machine cluster like Strix Halo or Nvidia Spark).
That means that to make the model fit into memory there you need a quantisation of lower than 4bits per weight (which is usually bad) to fit it into the available memory.
Tomte 3 hours ago
If only they managed to tell the mobile app to tell the model to reply in English to English prompts.
I suffix everything with "Reply in English", and even so I‘m getting lots of Chinese.
danielspace23 2 hours ago
I think their system prompt is in Chinese and probably has instructions to prioritize answering in Chinese, since this has never happened to me via API, where I (or the coding harness) set the system prompt.
monster_truck 2 hours ago
I just started learning Chinese instead, like they want us to
seriously
orbital-decay 2 hours ago
English isn't the first language for me as well so I don't see any problem with that
Grimblewald 3 hours ago
I'm starting to have chinese characters bleed into claude as well. Perhaps a sign of the times. Understanable for a chinese first model but an english first (supposedly) model? wild stuff.
donquichotte 2 hours ago
I also love the gaslighting of some models, like ChatGPT mixing in words with cyrillic letters and when asked about it answers: "it can look as Slavic to the eye" and "sorry that it came across as Russian"
calgoo 2 hours ago
Yes, this is one of the few issues with Deepseek; their chat pages and the app all respond in Chinese. However, i think i have only had it happen once when using the API, and im using it for hours each day for the last... couple of months?
SSLy 33 minutes ago
last couple of weeks, before they've unified instant and expert the former always replied in chinese unless steered, expert was by default english
sschueller 2 hours ago
Same issue on desktop. Would be nice be able to set a prefix or postfix for every prompt.
Markoff 44 minutes ago
nothing to do with mobile app, I have same issues while using it on desktop browser, it will never remember to use English permanently, even within one conversation
ignoramous 2 hours ago
I occassionally get Chinese characters interlaced with English in Google AI Mode, too.
swiftcoder an hour ago
OpenCode Go is currently running a 4x usage promo on DeepSeek v4.1 flash, not a bad way to get your feet wet (even if their cache hit prices are probably still very sub-optimal)
cdnsteve an hour ago
Hit me up if anyone wants extra $5 free usage with my referral code
RockstarSprain 37 minutes ago
Never tried OpenCode Go so I am interested. How does their pricing compare to paying DeepSeek directly, by the way?
cdnsteve 24 minutes ago
DavCreator 3 hours ago
jimmyl02 3 hours ago
The architecture changes and systems improvements being brought into LLMs is so awesome to see. It really feels like this is now a systems problem where a defined goal is set then systems optimizations are made around the model architecture to solve it.
Underlying it all is that any architecture can be trained to the same convergence just difference in compute utilization both in training and inference
bhouston 3 hours ago
Yes, this is called RSI, e.g. recursive self-improvement. It is the current stage of things and it is part of a hard takeoff.
lionkor 3 hours ago
I'm a big fan of DeepSeek. Also, ask it what model it is :)
In Pi (pi.dev), it tells me it's definitely Claude by Anthropic, via the API via curl it tells me it's "probably ChatGPT", its very funny.
Mashimo an hour ago
Works correctly in opencode, but seems like they inject a system prompt:
Thinking: > The user is asking what model I am. According to my system prompt, I'm powered by "deepseek-flash" with model ID "opencode-go/deepseek-flash".
>I'm powered by the model opencode-go/deepseek-flash.
shunia_huang an hour ago
Definitely not Claude, deepseek is too fast, so I bet it's ChatGPT. :P
gkbrk 23 minutes ago
Official Deepseek v4.1 Flash API costs are more than GPT 5.6 Luna. Deepseek v4 Pro performed worse than Luna, so I wonder if 4.1 Flash will justify the cost.
walrus01 2 hours ago
Looking at the huggingface page, the unsloth people haven't finished quantizing it yet, but I'm sure they're active on it right now. It'll be interesting to see how the capabilities and benchmark tests compare on system where it can fit in under 512GB of RAM with full context.
In terms of coding and command line capabilities I'm also very interested to see a head-to-head of it vs. qwen 3.8-flash-next Q8 which is something like 190GB of memory used when loaded into llama-server. It fits very well in all sorts of 256GB or under class machines.
k__ 3 hours ago
So, while the throughput was 400-500tps in beta its now ~150tps on OpenRouter.
I was hoping for a bit more, but it's still 100% faster for a very good price, so I won't complain.
gosolozero 3 hours ago
First flash model with multimodal support? I think Flash series might be the main focus going forward for them. Tried it out and it’s better than v4 pro
lionkor 3 hours ago
v4 pro is being discontinued, pasted the email here: https://news.ycombinator.com/item?id=49639667
thefossguy69 an hour ago
Makes sense. The 0731 snapshot of V4-Flash really made reaching out to Claude really infrequent for me.
arjie an hour ago
No. DSv4-Flash-Vision-Exp is what I use and it has vision.
viktorcode 30 minutes ago
It is now redirected to Flash v4.1
Alifatisk an hour ago
> New Causal Encoder–Decoder architecture: just 8B active parameters for input, 16B for output.
Oh interesting, I can assume what the benefits is for including the Encoder, but whats the downside? I’m thinking GPT (which is decoder only) ruled out Encoder for a reason?
Alpha3031 31 minutes ago
Enc-decs are usually harder to train at frontier scale. Not 100% sure what DeepSeek has done differently here initial read seems to be something related to layer reuse but I just skimmed things so far.
NitpickLawyer 3 hours ago
Jesus, this is a whole nother beast, and a different architecture from their previous flash. Lots of goodies here.
> Causal Encoder-Decoder (CED) architecture: a 40-layer Transformer organized as a 20-layer causal encoder followed by a 20-layer decoder. With CED, the decoder's global KV cache is projected from the final encoder hidden states rather than derived from each decoder layer's own hidden states. This allows the model to activate only 8B parameters per token during prefill and 16B during decode, substantially improving cost efficiency for input-heavy agentic workloads.
> these designs reduce the global KV cache footprint to 890 bytes per token — roughly 1/4 of DeepSeek-V4-Flash.
Faster prefill, lower kv cache (~1GB / 1m context is insane).
> The model supports a continuously controllable reasoning effort setting (integer 1–100) that trades inference cost for accuracy.
Benchmarks are benchmarks, to be seen if they translate to real-world use, but they seem to have focused a lot on post-training with "agentic" scores looking good. "world knowledge" is obviously lower than higher param models.
a012 3 hours ago
Waiting this model to be on openrouter (with other providers) to test out. In my use case, the GLM 5.3 Flash is the current cheapest and intelligent Flash model, but it’s dog slow at 13tps so I have to leave it run for many minutes then check again then correct it again
drob518 an hour ago
The speed of GLM 5.3 Flash on OpenRouter seems to vary considerably by provider. Some are fast and some are slow. OpenRouter does provide some tuning knobs, but not enough for my taste. It’s also token-heavy with reasoning, though I found it better than Deepseek V4 Flash previously.
shunia_huang an hour ago
> though I found it better than Deepseek V4 Flash previously
Same experience here.
But man, switch to V4.1 now! It is much better.
I don't event need to test it for long run and I believe it's crazy good. I call it "AI era model taste" when I judge the model by it's output without reading the bench scores.
karimf 2 hours ago
While this is very impressive benchmark-wise, GPT-6 Astra showed us that benchmarks don't always correlate 1:1 to intelligence of a model.
When Astra launched, I think Artifical Analysis showed that it was on par with GPT-5.6 Sol and lower than Opus or something like that? Then, they updated the scoring.
I hope that more open source models, including this model, to be "as good to use" as Astra.
Squarex 2 hours ago
I don't know why, but the benchmarks still fails to cover the difference between large models and small ones. The small ones are great for many things, including general coding, but the larger ones, like fable and astra, have some kind of intelligence that is not present in the small ones.
sinuhe69 an hour ago
More parameters = more facts stored. Knowledges are almost incompressible, where strong reasoning only requires a 3B core or so.
yorwba 34 minutes ago
walrus01 2 hours ago
Apparently the scoring on a lot of difficult benchmarks can also be extremely influenced by something as simple as waiting for the model to exhaust its reasoning, realize it hasn't come to a conclusion yet, and give it a simple prompt like "you can do this, I know you're capable, please keep going".
raesene9 an hour ago
This seems like a very nice release. Just ran it over my Kubernetes security benchmark that I run for most new releases. It was fast, cheap, and got a high scoring result, nice!
E-Reverance 4 hours ago
The figure on page 5 in [1] is pretty insane
[1] https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/...
schneehertz 4 hours ago
A very powerful model, and with multimodal support now, it can be used as a primary model.
WalterGR 3 hours ago
Related: https://news.ycombinator.com/item?id=49624603
“DeepSeek launching v4.1 flash cheaper and more capable than v4 pro”
399 points | 19 hours ago | 216 comments
theanonymousone 2 hours ago
For technical report: https://news.ycombinator.com/item?id=49639110
arj 2 hours ago
Having this available to find and fix security stuff is a big deal. The model of really good.
SyneRyder 2 hours ago
Just a reminder that if you want to try this via OpenRouter, DeepSeek openly trains on all of your prompts. So maybe don't go using this to solve the last unforced step of Navier-Stokes. (Or wait until some other providers start hosting this with ZDR or other policies, which shouldn't be too long.)
Lucasoato 2 hours ago
My question is: what kind of hardware do you need to run this Flash beast locally at a meaningful speed?
aenis an hour ago
8x RTX PRO 6000 or 4x Spark? Or 1x M5 Ultra 512GB.
The model is theoretically FP8, but really internally its mostly FP4 already, so there won't be a cut-in-half-but-almost-just-as-good quant coming for this one.
ekianjo 2 hours ago
a beefy pc with at least 20 GPUs
linzhangrun 2 hours ago
They say v4.1flash is so strong that they'll route API calls to v4pro to v4.1flash, lol
super fast true
lwansbrough 2 hours ago
Significant jump in pricing. V4 Flash was $0.16/M out, 4.1 is $1.20/M.
svantana an hour ago
I think you're comparing to third party prices, deepseek's prices hasn't changed with this release. Also, $1.2 is the peaktime price.
trq01758 2 hours ago
Never saw $0.16 for 1M output tokens - it was $0.28 a month ago, $0.66 off-peak and $1.32 peak last week, now it is reduced a bit to $0.6 and $1.2
lwansbrough 12 minutes ago
Was looking at OpenRouter, I guess it’s wrong.
dakolli 2 hours ago
incorrect, no idea where you're getting this pricing. Also, output does not matter. its 10% of the cost.
bertili 2 hours ago
The bigger story is the compute efficiency - its been running at 300t/s the last days.
arjie an hour ago
What in the world. A point release with 2x the parameters and a different architecture? Jesus. Can’t run this kind of thing on 2x RTX Pro 6k at decent speed. I need to reconfigure my hardware. Massive disappointment on that front. Bloody hell. Glad I didn’t get a DGX Station.
No wonder they retired the Pro model in favour of this.
jonplackett 2 hours ago
Can we just never link to X posts as the main link.
nunodonato 4 minutes ago
yes, please. Especially now that xcancel is gone
thatsadude 2 hours ago
DeepSeek invented the whole reasoning paradigm and keep pushing for innovation. I hope they get the success they deserve.
yorwba an hour ago
OpenAI released their first reasoning model (o1-preview) https://openai.com/index/introducing-openai-o1-preview/ several months before DeepSeek's R1 https://arxiv.org/abs/2501.12948
_davide_ 3 minutes ago
CoT, was being studied using GPT-2, so...who invented hot water first?
mohsen1 3 hours ago
I am speculating but hard to not see that DeepSeek is brewing a full Pro model with those new techniques to come out right around the time of Anthropic and/or OpenAI IPO to tamper the excitement for their offering.
ignoramous 2 hours ago
DeepSeek will deprecate the v4 Pro model (it will route to v4.1 Flash starting 14 Sep). Unsure what comes next, but I'd wager a bigger model à la Kimi K3: https://news.ycombinator.com/item?id=49639667
siscia an hour ago
I am building software factories and deepseek IS the workhorse.
I personally found V4-flash an amazing model and really hungry to try 4.1-flash
For software factories, cost is much more a concern that standard development workflow and using anthropic models is just a non starter