Hacker News

points by numlock86 3 days ago | 0 comments

> Meanwhile the Chinese are going to undercut them every chance they get.

I wish they were, but I always wonder when people chime in with Chinese models what they are actually doing with those. I have an OpenRouter account and give various models a shot every other month, wishing for an alternative, but I always come back to Claude. Even more so since the switch from Opus 5.1 to 5.5. To me, personally, it's unmatched. Like a year ago it was very different, and I would use other models all the time. But honestly at the time they all felt kind of "meh" still. I just hope everyone is at a good baseline in a year from now on, so I actually have a choice. And then, as you stated, things will get interesting.

tl;dr: I agree

[–] jacquesm 3 days ago

> I always wonder when people chime in with Chinese models what they are actually doing with those

In my case: robotics, complex document analysis (typically 800K context).

[–] Sphax 3 days ago

I use Opus 5.5 daily at work and a mix of GLM 5.3 (Flash or not), Deepseek v4 (flash or pro) at home and while Opus 5.5 is more capable, in my opinion it matters for less than 5% of the work i do. These chinese models are simply good enough for the vast majority of tasks I run and when they're not I use Astra or Opus 5.5. Yes, leaving Opus 5.5 on all the time is comfortable when you pay the 90€ plan but it's just not necessary.

[–] criley2 3 days ago

I use Opus 5.5 at work and I use Sonnet 5.5 at home. Sonnet 5.5 medium according to AA cost per task is $0.48 per task, while GLM 5.3 (max) is $2.01 (opus 5.5 high is $1.82). GLM 5.3 flash is $0.28 per task and Deepseek 4.1 flash 1017 is $0.26.

So it seems like Sonnet 5.5 medium, GLM 5.3 flash and Deepseek V4.1 flash 1017 are all in a similar space in capacity and price, however the $20 anthropic sub will get you like $200-$300 in sonnet 5.5...

Another bonus is that sonnet uses a fraction of the overall tokens and finishes much faster than the average chinese model, so your work goes faster (Sonnet 5.5 medium in 2 minutes, Deepseek 1017 in 6.5 minutes, GLM 5.3 flash in 13 minutes...)

I still don't get why paying for chinese inference makes sense. I should try Haiku 5.5 high/max because it's posting higher numbers than Deepseek V4.1 flash for half the price

[–] ac29 3 days ago

Yeah, I spent the last week with Deepseek Flash 4.1-1017, Kimi 3, and GLM 5.3 (and 5.3-flash). None of them were literally useless but also none of them were capable of making progress on the main project I am working on where Opus, Sol, and even Gemini Flash have all done very good work. Benchmarks might have you believe open models are only barely behind the big US labs in intelligence, but it feels much worse in practice.

[–] jacquesm 3 days ago

Interesting! If you are at liberty to disclose, what problem space are you working in, what harness do you use?

[–] joshheitzman 3 days ago

As a counter-example, I use Deepseek v4.1 Flash to make progress on multiple projects (Python, C++, CUDA) every single day.

[–] maupin 3 days ago

I am running GLM 5.3 Flash locally at home for coding and from my point of view it is indistinguishable from the frontier models of 6 months ago. It can handle almost every task I throw at it.

[–] creativeSlumber 3 days ago

What is your setup like? How much VRAM do you have? I'm curious about what is the minimum configuration that you would need to run this locally.

[–] maupin 3 days ago

I have 2 DGX Sparks, 256 GB unified RAM, so not a cheap set of hardware. Luckily I bought them before the recent price increase.

However even with a high tier gaming machine you can run one of the Qwen models and it will be good at coding... like having an intern with boundless energy that will sometimes make mistakes but in general save you a lot of time. And for automating tasks involving scanning documents or creating summaries it will be brilliant.

[–] jacquesm 3 days ago

Not the GP, but in my case 240G VRAM and 256G RAM will run the 3.86 bpw version with a little bit of room to spare and a very nice size context.

[–] motbus3 3 days ago

i reduced my claude expenditure in 80% using other models.

[–] emtel 3 days ago

I don’t know why you’re getting downvoted. One of my coworkers who was constantly touting Chinese models a few months ago now only uses opus 5.5. It is a huge difference.