Hy3
111 points
1 hour ago
| 12 comments
| hy.tencent.com
| HN
simonw
29 minutes ago
[-]
Pelican from a few days ago: https://simonwillison.net/2026/Jul/6/hy3/ - I was using the free tier on OpenRouter, which expires on July 21st.

I tried the preview model 41 days ago and got a pelican with a "change pelican color" button: https://static.simonwillison.net/static/2026/hy3-preview-pel...

reply
minimaxir
44 minutes ago
[-]
A month ago I wrote a blog post about how Hy3 was topping the OpenRouter rankings despite no one talking about it: https://news.ycombinator.com/item?id=48317294

As of today, it has fallen to 8/9th on the rankings. I don't see a reason where you would use this model over competitors. However, price economics are bit confusing, as currently the effective input price of Hy3 via OpenRouter is now the same as DeepSeek-hosted DeepSeek Flash V4.

https://openrouter.ai/tencent/hy3-preview

https://openrouter.ai/deepseek/deepseek-v4-flash

reply
Miner49er
35 minutes ago
[-]
I had to stop using it because I was getting rate limited like crazy. Probably why it has dropped. Seemed like they couldn't keep up with demand.
reply
docheinestages
4 minutes ago
[-]
What we really need is a breakthrough in inference or LLM architecture to allow running GLM-5.2-level models at the size of Qwen 3.6 27b or smaller on consumer devices like a 48GB Macbook Pro, and at least at 100 tokens/second. My hypothesis is that a smaller, less capable but faster model paired with a good harness can run for longer and brute force its way out to solve problems that the bigger models can one-shot.
reply
Catloafdev
1 hour ago
[-]
Curious how people feel about this compared to DS4 Flash, given they are pretty close in size. Also curious how well it holds up to heavy quantization.

DS4 Flash can currently run reasonably well on systems with ~96gb+ RAM, I wonder if Hy3 can compete there.

reply
tarruda
12 minutes ago
[-]
> given they are pretty close in size

One thing that might not be obvious about about DSV4 is how much innovation the Deepseek team implemented in its architecture. When llama.cpp fully supports its lightning indexer, the full 1M context will only require about 6G of RAM. So even though they are similar in size, I believe Deepseek will be much more efficient in that regard.

> I wonder if Hy3 can compete there

Highly depends on how well Hy3 is resilient to quantization. DSV4 is useful even at 2-bit quants.

reply
UncleOxidant
1 hour ago
[-]
That's a 2-bit quant of DS4 flash. You're probably better off running Qwen3.6-27B at Q8.
reply
spmurrayzzz
49 minutes ago
[-]
I think its good advice to test both on your own evals for sure, but the MoE parameters are already natively FP4 in ds4. Dropping to 2bpw isn't as big of a loss as it seems (and as corroborated by antirez's work).

Its also only 13B active, so your decode speed would be nearly 2x that of Qwen3.6-27B. So there are other latent benefits as well.

reply
Catloafdev
34 minutes ago
[-]
For most coding or agentic tasks, Qwen 3.6 27B likely outperforms, yes.

For 'general intelligence', DS4 Flash seems to be a noticeable step up still.

reply
sosodev
55 minutes ago
[-]
I suspect it would depend on the task. DS4-flash does, as previously mentioned, handle quantization very well. Even at 2-bit it's still very coherent.
reply
ckocagil
22 minutes ago
[-]
Isn't Q8 way overkill these days? I see many graphs showing Q4 or Q5 having less than %1 deviation. Nvidia's NVFP4 Qwen quantization should be even better due to its better training methods.
reply
mdgld
20 minutes ago
[-]
It depends on model size I think, but yeah, from my understanding at ~30B and below Q6 or even Q4 will get you 95%+ of the way there
reply
wolttam
1 hour ago
[-]
Hy3 lacks the DSv4 architecture's KV Cache efficiency.

Whereas I can run DSv4 Flash on a pair of DGX Sparks and have enough memory left over for 3M tokens of KV cache, with Hy3 (quantized to FP4), there is only room for ~130K tokens of KV cache.

reply
ignoramous
48 minutes ago
[-]
Lower context window notwithstanding, Hy3's coding benchmarks hold their own against DeepSeek v4 Pro & MiMo v2.5 Pro. That's quite something for a model priced like DeepSeek v4 Flash & MiMo v2.5 (for non-cached tokens), which are 3x cheaper than their respective Pro variants.
reply
wolttam
41 minutes ago
[-]
It's impressive indeed. I would also expect the next checkpoint of DSv4 Flash to come in somewhere at this level (DeepSeek has had over 2 months to continue training since it released).

It's exciting that the open models continue to get better and more efficient across the board!

reply
nunodonato
1 hour ago
[-]
DS4-Flash is not only "significantly" smaller, it will also benefit from a lot more speed thanks to DSpark
reply
Catloafdev
59 minutes ago
[-]
299B for Hy3 vs 284B* for Flash

Edit: fixed, got bad info

reply
nshotton
42 minutes ago
[-]
reply
Catloafdev
38 minutes ago
[-]
Oh, it is. I was looking at the Huggingface repo which listed the lower number at the top of the page, looks like that's wrong.
reply
thot_experiment
27 minutes ago
[-]
I feel like I'm taking crazy pills with hy3, it's either benchmaxxed to hell and back or skill issue on my part but I'd rather use dense gemma. I don't think there's a single model that's wasted more of my time in recent memory.
reply
nshotton
1 hour ago
[-]
This model is shockingly small for how capable it is. its a little bit bigger than deepseekV4 flash but around as capable if not more on some benchmarks than V4 pro, i wouldnt be surprised if this becomes a popular local model.
reply
andai
1 hour ago
[-]
I've been wondering about that. GLM-5.2 is also half the size of DeepSeek V4 Pro. (But costs roughly twice as much.)

I looked into DeepSeek's architecture a little bit and the main focus was how can we save as much money as possible. They did a lot of cost cutting with the attention mechanisms. This allowed them to offer an insanely cheap price even on massive contexts, but seems to have come at the cost of performance?

At least, that's my guess, when I see smaller models costing more and outperforming, I think, "they must have denser attention?"

reply
wgd
28 minutes ago
[-]
The current Deepseek V4 Pro is still just their initial preview AFAIK, with the "real" model release rumored to come later this month. GLM-5.2 might be outperforming simply because it's had more post-training on top of the GLM-5 base.
reply
nunodonato
1 hour ago
[-]
hardly, its still quite big unless by "local" you mean people that spend many thousands on rigs :)
reply
nshotton
1 hour ago
[-]
Yeah i shouldve been more clear, a model of this size could run on 2 dgx sparks so out of the range of a lot of the typical consumer sure, but I think there is definitely a market for that size
reply
IshKebab
1 hour ago
[-]
> Hy3 has 295B parameters in total. To serve it on 8 GPUs, we recommend using H20-3e or other GPUs with larger memory capacity.

I would.

reply
minraws
1 hour ago
[-]
I tried out the model it's pretty great, better than ~~gpt5.4~~ gpt-5.4-mini perhaps, atleast close enough to sonnet 5 in performance that I didn't notice much of a gap.

Not really at gpt 5.5 tier though, and probably below glm 5.2...

But most of all it just works for me for most things I tried and it's exceedingly cheap so there is no reason not to use it, if you need a foss model.

Edited: gpt-5.4-mini not the base gpt-5.4

reply
theplumber
53 minutes ago
[-]
I think you’ve got the models wrong…gpt-5.4? I doubt there is any open source mode matching it. Maybe in a year
reply
minraws
46 minutes ago
[-]
Yeah I meant gpt-5.4-mini, but GLM 5.2 is pretty close to gpt-5.4 base, and much better than it when it comes to design stuff.
reply
mgrandl
44 minutes ago
[-]
GLM 5.2 already matches GPT-5.4 easily.
reply
cbg0
53 minutes ago
[-]
Hy3 DeepSWE - 28%

GPT5.4 xhigh DeepSWE - 52%

A lot of contaminated benchmarks in the blog post about Hy3, needs real testing though I have a distinct feeling it's benchmaxxed like a lot of Chinese models.

reply
throwaway2027
1 hour ago
[-]
Quite interesting to see them and Meta and others release before OpenAI supposedly is to release GPT 5.6 today, would it be better to release it before or after? Calm before the storm type of thing?
reply
james2doyle
1 hour ago
[-]
Been using this and GLM 5.2 back and forth. I like the speed of Hy3. Also seems very happy to follow instructions. Still haven’t found any open models that follow instructions as good as Mimo v2 pro though
reply
handzhiev
1 hour ago
[-]
It's a very good model for this size and price. I tried it with a couple of small tasks - just an year ago this would be the level of the leading models.
reply
doawoo
40 minutes ago
[-]
That UI demo page is… really quite janky.
reply
nunodonato
1 hour ago
[-]
Very impressive model for its size
reply