News for hackers. Stuff that matters.
Re:DeepSeek-V4-Flash Update by f311a on Friday July 31, 2026 @03:46AM (#49120203)
I've been driving flash model for 90% of my tasks. It's better than pro (for unknown reasons), very cheap and fast.

I try to keep changes under 1000 lines and drive architectural decisions myself, barely notice any difference compared to frontier models. The rest 10% is to spot bugs, security problems and to investigate better architecture, which flash can also do pretty well, I just cross check it.

Faster iterations are way better for me, I hate waiting for 5-10 minutes on small changes. I tried to use recent versions of Kimi and GLM, but they use too much thinking for no reason and are pretty slow because of it. I also often feed a lot of data to it, without worrying about hitting the limits: dependencies (to find bottlenecks in them), logs, performance dumps and so on.

Also, it will never complain about security guards, I've been using it to reverse engineer binaries.

Re:DeepSeek-V4-Flash Update by regularfry on Friday July 31, 2026 @04:02AM (#49120315)
It's good but (at least on openrouter) it's got an annoyingly tight output token limit. So if it does get stuck in a reasoning pit, it won't work its way out of it in time.

It's replaced the Kimi models for me though.

Re:DeepSeek-V4-Flash Update by u8080 on Friday July 31, 2026 @04:44AM (#49120574)
Try to use DS platform directly - cheaper and better than openrouter, no subscription
Re:DeepSeek-V4-Flash Update by dghlsakjg on Friday July 31, 2026 @10:46AM (#49123863)
Openrouter has cheaper inference than deepseek for the prior version of v4 flash, which I suspect will happen within a day or two with this version, and they don’t offer subscriptions. Are you confusing it with opencode?
Re:DeepSeek-V4-Flash Update by u8080 on Friday July 31, 2026 @01:19PM (#49126021)
No, maybe I've checked too much time ago, but direct DS was cheaper for all versions - also I had issues with openrouter providers availability
Re:DeepSeek-V4-Flash Update by dghlsakjg on Friday July 31, 2026 @02:09PM (#49126668)
There are numerous providers offering full weight versions for a significant discount.

Availability and rate limiting can be an issue, though. I’ve found that constraining the providers works well to solve that though.

Re:DeepSeek-V4-Flash Update by embedding-shape on Friday July 31, 2026 @04:24AM (#49120459)
> Also, it will never complain about security guards, I've been using it to reverse engineer binaries.

Maybe I'm using too weak language in my prompts, but none of the OpenAI models I've used via codex has refused to reverse engineer binaries, is it supposed to? I'm sitting right now reverse-engineering a 3rd party firmware together with Codex and haven't hit a single guardrail. Meanwhile, I see people complaining about it rejecting non-security related prompts, are things so individual on the platforms right now or what's going on?

Re:DeepSeek-V4-Flash Update by markasoftware on Friday July 31, 2026 @04:42AM (#49120564)
Have you completed the identity verification? It's much more lenient once you have
Re:DeepSeek-V4-Flash Update by embedding-shape on Friday July 31, 2026 @04:44AM (#49120577)
Oh yeah, back in the GPT3.5 days I think, that's probably it. Thanks for sharing your hunch :)
Re:DeepSeek-V4-Flash Update by flexagoon on Friday July 31, 2026 @04:51AM (#49120628)
Weird, I'm using 5.6 Sol through Chinese resellers and it reverse engineers stuff just fine
Re:DeepSeek-V4-Flash Update by ngl999 on Friday July 31, 2026 @05:40AM (#49120989)
Probably not real 5.6 Sol
Re:DeepSeek-V4-Flash Update by flexagoon on Friday July 31, 2026 @07:10AM (#49121668)
How so? The code and reasoning patterns certainly match, so does the intelligence level. I probably wouldn't know if they're secretly serving Luna instead of Sol, but it's definitely an OpenAI model.
Re:DeepSeek-V4-Flash Update by genxy on Friday July 31, 2026 @01:13PM (#49125947)
Because they have gone through identify verification, they are more likely to have done that to increase reputation score.
Re:DeepSeek-V4-Flash Update by dotancohen on Friday July 31, 2026 @01:42PM (#49126319)
Do you have any suggestions for resellers? My Gmail username is the same as my HN username if you're not comfortable posting that here.

Thanks.

Re:DeepSeek-V4-Flash Update by markasoftware on Friday July 31, 2026 @04:39PM (#49128278)
Re:DeepSeek-V4-Flash Update by dotancohen on Friday July 31, 2026 @05:07PM (#49128529)
Thank you
Re:DeepSeek-V4-Flash Update by flexagoon on Saturday August 01, 2026 @08:44AM (#49133974)
I used Byesu and now use a6api. Byesu is a bit more expensive, but a6api is a marketplace so has a higher risk of fake models.
Re:DeepSeek-V4-Flash Update by OkGoDoIt on Friday July 31, 2026 @01:46PM (#49126365)
I think I’m missing something, what do you mean Chinese resellers? As I understand it, it’s difficult to even access OpenAI in China, how could they be reselling it? Do you mean something like openrouter but a Chinese version or something?
Re:DeepSeek-V4-Flash Update by Scoundreller on Friday July 31, 2026 @03:13PM (#49127394)
Proxies/vpns to enter/exit comms through non-banned countries.
Re:DeepSeek-V4-Flash Update by markasoftware on Friday July 31, 2026 @04:38PM (#49128263)
They usually resell codex subscriptions as api so it's cheaper than the official api
Re:DeepSeek-V4-Flash Update by johntarter on Friday July 31, 2026 @02:52PM (#49127183)
That's hilarious that Anthropic and OpenAI can't even secure their products from Chinese resellers, how are they supposed to secure their models from being distilled?
Re:DeepSeek-V4-Flash Update by vablings on Friday July 31, 2026 @05:40PM (#49128832)
They can't, that's the entire market for resellers all the data collected is used for distillation
Re:DeepSeek-V4-Flash Update by econ on Saturday August 01, 2026 @01:53PM (#49136699)
Same way you prevent them from crawling your public website?
Re:DeepSeek-V4-Flash Update by realusername on Friday July 31, 2026 @04:43AM (#49120573)
I got an account warning on OpenAI (waved after I complained) just because I was asking it how to root some >10 years old Android device.
Re:DeepSeek-V4-Flash Update by dotancohen on Friday July 31, 2026 @01:44PM (#49126346)
Sonnet recently even suggested I root a three year old device and provided instructions. I suppose that is depends on use case. My use case was exporting data from an abandoned application running on an S24 Ultra.

The idea was that I could continue to use the application in the future and export the new data, not that I would be able to recover the extent data already in there.

Re:DeepSeek-V4-Flash Update by mark_l_watson on Friday July 31, 2026 @07:28AM (#49121814)
This is my experience also: DeepSeek v4 flash is good enough for most of my work and I like the fast response times. I buy tokens mostly from FireWorks.ai in the US, but I also prepaid for a large chunk of tokens directoy with DeepSeek.

I use OpenCode mostly (uses fewer tokens than Claude Code) and I am looking forward to the release of DeepSeek’s own coding harness.

Re:DeepSeek-V4-Flash Update by quicekuru on Sunday August 02, 2026 @07:26AM (#49143426)
Same experience here. Plugged it into my harness instead of a 10x more expensive model and it... just worked, sometimes even better than the expensive model in a direct comparison. Starting to think there is more value in speed than "intelligence" at this point.