OpenRouter is a place with zero support. Suddenly get a big debt on your account with nobody to respond. As an early adopter of OpenRouter, I'm afraid they are in shambles.
Isn't that the model for all of the labs though?
DeepSeek V4 Pro 0813
471–480 of 493 posts
Re: DeepSeek V4 Pro 0813
#472Nice bicycle chain, the little basket with a fish didn't show up in the right place: https://tools.simonwillison.net/markdown-svg-renderer#url=ht...
Might be time to also have it try more three dimension pelican rendering. Or a short animated version (even just a few frames) still in standard SVG
Re: DeepSeek V4 Pro 0813
#473Currently burning money quickly on official deepseek api. They are also increasing pricing starting today. V4 Flash 0731 still feels like the most outstanding model of the past few months and probably to come.
What is the new price through?
It has been updated now, and will take effect next week:
https://api-docs.deepseek.com/quick_start/pricing/ Briefly Pro is $2/1m output in off-peak periods, $4 in peak. Flash is $0.66/$1.32. Input tokens are still much cheaper.
I don't mind these prices but I find the need to check against two different time brackets of unequal length an annoying distraction. I guess I need to make some little background app or plugin.
Re: DeepSeek V4 Pro 0813
#474[flagged]
Re: DeepSeek V4 Pro 0813
#475Earlier quoted context omitted.
An agent doing a task even with multiple back to back calls like normal without an example is zero shot. An agent doing a task with 1 example is one shot. An agent doing a task with a few examples is few shot. I don't think you are correctly using these terms. The multiple back to back LLM calls are done on accumulating context, so if there is a sampling error it could throw the entire session out of whack, because L…
> An agent doing a task with 1 example is one shot. An agent doing a task with a few examples is few shot. I don't think you are correctly using these terms This is a different thing. Yes, giving multiple example is called "few-shot prompting". But one-shot vs few-shot benchmarking is different. In this context "one-shot" means "pass at 1 effort" as opposed to "multi-shot". In the literature this is called "pass@k".…
Yes, and the reason why pass@k exists is because of self-consistency. There is no guarantee for right answer to be selected or for the LLM to correct itself. While I agree pass@1 is a useful metric, I'd be more interested to know pass@5 so I can better compare the results.
Re: DeepSeek V4 Pro 0813
#476Nice bicycle chain, the little basket with a fish didn't show up in the right place: https://tools.simonwillison.net/markdown-svg-renderer#url=ht...
For a while now, I've found pelican rendering to be an unreliable metric for LLM ability - and most people know it. Yet, somehow it gets upvoted to the very top of every new model discussion.
Because seeing a pelican on a bike is always a good time. Look at it go
Re: DeepSeek V4 Pro 0813
#477Earlier quoted context omitted.
Yeah pretty much. I spent half a billion in tokens one night on a huge refactor with DSFlash, cost $11. If I spent that every night it would be 3x my GPT subscription.
That seems too expensive to be honest. Did you do it with official deepseek api or a 3rd party provider? Because official has 10x cheaper cache reads than the rest. I've done similar sized chats for like $1
That pricing is going away soon. It's about to get 5x more expensive.
Re: DeepSeek V4 Pro 0813
#478Re: DeepSeek V4 Pro 0813
#479Earlier quoted context omitted.
You accept any and every registration? I think that goes a bit beyond an open market and into a completely unregulated one I'd never want to route my sessions through. Thanks for providing the conflicting interest disclosure though, too often people don't bother.
No, not any
So why are rejections only a problem for the other marketplace's openness?
Re: DeepSeek V4 Pro 0813
#480Earlier quoted context omitted.
With confidential compute / TEEs you can guarantee that the code is running, it's verifiable with remote attestation
So where does the guarantee stop? At the GPU driver level? Firmware level? What if the GPU has a custom bios flash that somehow logs the unencrypted prompts?
goes all the way to keys owned by Nvidia and Intel