Join the discussion
Write your take first — we'll ask for email only when you're ready to publish.
- Hacker News
- If they can keep up this cadence of Flash leap-frogging the previous Pro, we're in for a good timeby swiftcoder
- Anthropic/OpenAI might step up their anti-distillation defences though.by dude250711
- I've been watching a bunch of bycloud on YouTube recently, and although he's done a great job reviewing papers from the big AI labs, I feel like I'm missing something - how have all the labs seemingly made a model that's cheaper, faster AND has better performance? Historically `flash` variants (like codex spark as well) have been faster but perform worseby edude03
- That’s how increasing performance works. You make a model 10x faster, then you make it think 2x as much.
Its cost is now 1/10th per token, and 1/5th per task.
Basically they have shitty hardware so they have to do a lot of optimization. Think of it like replacing an O(n) algorithm with O(log n).
Anthropic / Open AI think the best path is the most intelligent models deepseek is more focused on tok/$
by _3u10 - It's interesting that this is the third lab to find problems with larger models. Earlier last year oAI was rumoured to have failed their large pretrain. Now google has problems with their pro series, and ds just announced the same. There are some rumours on chinese forums talking about problems with the pretraining phase, so this is not mid/post training related.
I wonder if this comes from using the bad architecture scaled up (and it hits some limits) or if this is a data problem (undertrained? bad data? bad pre-processing using smaller models?)...
- Where are you seeing them having an issue with the larger (Pro) model?
The announcement specifically says 4.1 Pro will be released in the future.
by pixelesque - Just my intuition about it but it does seem like a data issue.
V4 flash and V4 pro feel very similar, which would make sense if they were pre-trained on largely the same corpus.
All that would suggest to me is that V4 Flash is capable of absorbing the data they’re throwing at it, and we’re still nowhere near the data limits of their larger 1.6T model
by wolttam - Hopefully it will be open weights and have the same architecture and size as the current v4 flash vision, which is probably the best LLM that can be run on 128G devices.by tarruda
- Interesting, I had assumed it'd be too large to fit. What quant and context size are you running?by fluoridation
- Beta testers report >400 TPS.
https://www.geeky-gadgets.com/deepseek-v4-1-flash-review/
I hope some of those speed increases will make it to production.
by k__ - That ought to be DeepSeek's real differentiator; all the other Chinese models are slow.by esafak
- I will continue to be amazed by how much power you get from DeepSeek Flash for the cost. I have let that puppy lose on so many projects and it is has never let me down. It can build and entire Rails app in no time and even do the tests. For most things, I don't get why people pay the money for Claude. DeepSeek Flash is my default agent in Omarchy.
- How does it compare to GLM-5.3 Flash?
- Same, very impressed with v4 flash. It has the right balance of cost and performance.by bwfan123
- I hope DeepSeek takes some time to improve their tuning for reasoning effort. Right now, there are only three reasoning efforts: low, high, and max.
For all intents and purposes, "low" is pretty much the same as turning reasoning off, and "high" is similar to "max". "High/max" performs way too much reasoning, takes forever, and causes costs to balloon. They need a proper "medium" setting.
I get it that they're probably focused on pushing performance right now, but the ergonomics of the model aren't great.
by postalcoder - I would rather have just 3 levels: low, medium and high.by tarruda
- I switched to GLM-5.3 flash on high for this reason. Too many "but wait" in the Deepseek-v4 reasoning.by iamniels
- Since a few months, I almost exclusively use the Chinese "flash" models for my needs. They are a joy and they cost pennies per answer. Great job.by EbNar
- Do you use them for coding with your harness or do you use them in production? I found the latency distribution on OpenRouter to be unusable for DeepSeek v4 Flash.by iamjs
- I recently had to config my harness to watch for cybersecurity flags from astra and funnel requests to flash when they occur because Astra gets queezy when you talk to it about UDP packets in games.
Works fantastic. Glad there is a more 'uncensored' thing to fall back to when the frontier folk are too sensitive.
by serf - Same for me. DeepSeek models are incredibly good at implementation and light planning. I still default to Opus models for feature planning, but for most simple features the Pro models suffice.
Incredible good value and product they have built.
by XzAeRosho - My mental bias always kept me away from Chinese models. Because i know that china is a surveillance state and all the things we know about CCP. But after what we learned about OpenAI and how they most likely used user data to basically cheat in an open competition i think it does not matter which AI provider you use all of them will own your data and all of them can spy on you. So I am willing to switch to Chinese models. This way we help them develop and improve models some day we can run them locally.by darkoob12
- Yeah, the latest batch of <256GB Chinese models are really nice. They're far less cryptic than Claude, and competent enough to feel almost near Opus. I canceled all my subscriptions and switched to running the Chinese models locally (not as a cost saving measure).by hgoel
- Yes. I'm working in the agent industry and my god are we excited on new versions of Chinese flash models. The direct competition is Gemini Flash, and these models are much better on agentic tasks with fraction of the task price compared to Gemini. Things like oh here's a set of simple instructions for you to follow, call these tools, return this report. 20-30% of the price per task. And especially Deepseek Flash produces better quality than Gemini does.
Where Gemini still wins is non-text input what Deepseek cannot do, yet, and Deepseek Flash has this thing of cheaper models where a failing tool call can derail your agent to a retry loop if you're not careful on instructions in the error message.
If they fix and make the tool calls to work better in non-optimal situations, it's much easier to switch from Gemini without a few weeks of evals and bugfixing.
by pimeys - I am legitimately more excited for this release than any frontier models at this point.
I don't need a model that can invent new mathematics. I need something that is fast, cheap, and consistent. Give me that and I can build and scale.
by ActionHank - >In keeping with our commitment to user responsibility, following the official launch of V4.1 Flash and prior to the release of V4.1 Pro, all requests to the Pro model will be routed to V4.1 Flash and billed at Flash's price
Please don't do this kind of thing. If a user has validated a workflow on V4 Pro, they might not want to suddenly start testing it in production on V4.1 Flash. Instead, keep V4 Pro around but deprecated for a defined period of time, then remove it.
At least as open weights models, it's possible to use something like Together.ai or OpenRouter to run the V4 Pro model as long as other providers keep it up.
by aftbit - I wonder if this is a sign of things to come for dirt-cheap model hosting: no servers running old versions, only new versions. Just to keep costs down.by bicx
- I don't really agree, but we shouldn't have to debate it. An Auto option at each level would preclude this kind of decisioning. Pick a discrete model, that's what you get.
Pick Auto (Deepseek v4 Flash Auto vs Deepseek v4.x Flash), and let the vendor decide. I think OpenRouter uses this method.
by teamv02 - In this case, Deepseek organization is under a lot of pressure due to compute constraints. It would be better if they just throw a 404 instead of rerouting though so customers are not surprised by subtle changes in behavior.by samuelknight
- LLMs add enough nondeterminism to a workflow. Swapping them without the user knowing adds substantially more.by jmathai