Hunyuan Hy4 vs Qwen 3.8 Max
Comparar modelos
Hunyuan Hy4 is more worth it — 49% vs 43%
How we score worth-it? →Tencent's open-source MoE flagship — 770B total/49B active params, 1M context, top open-source tier for math and agentic work.
- Hy4 apparently has ludicrous traction on OpenRouter already, with trillions of tokens processed in a couple days. Relatively cheap with a 5% cache cost.— hn
- I'm very glad that models which are continuously improving in capability and getting closer to closed-source ones can be open-sourced.— huggingface
- WTF! That's insane. Someone test this lol.— reddit ↗
- The future of local models is brighter than expected. We can still compress them further — Hy4 has not yet been sufficiently post-trained.— reddit ↗
- And this is a PREVIEW? The actual release must be fable level or higher then?— reddit ↗
- Maybe it's a problem with the hosting, but I didn't get much useful out of this model as a coding agent.— hn
- I used CodeBuddy + Hy4 preview to explore a few things and it kept getting them wrong; switching to another model was the right call. I'll never trust Hy again.— v2ex ↗
- The only saving grace from an onslaught of 'it overthinks' posts is that it's a huge chonker of a model, too big for most people.— reddit ↗
- Aggressive quants don't degrade uniformly — short single-turn questions barely move, but long agentic runs fall apart around turn ten.— reddit ↗
- Thinking takes too long with Hy4 — it thought for over 10 minutes just to change a bit of frontend code.— blog ↗
“Horizon Math score is the best among all open models, second only to GPT-5.6 Sol.”
“This is a PREVIEW? The actual release must be fable level or higher.”
“Thinking takes too long — it thought over 10 minutes just to change a bit of frontend code.”
Alibaba's flagship 2.4T-parameter MoE model (95B active) with a 1M-token context, native image/video understanding, and strong agentic coding; first Max-class Qwen with open weights.
- Qwen3.8-Max has moved from 'knowing how to write code' toward 'being able to deliver tasks.' In structured engineering it shows a trustworthy prototype — good at breaking complex requirements into a project skeleton.— pingwest
- Per Bloomberg, Qwen3.8-Max costs $2 per million input and $6 per million output — output price is about one-fifth of some flagships, decisive over a long autonomous run.— pingwest
- The flat regional tier is the most concrete change to agent economics — a rate that holds from 5K to 900K tokens removes a planning variable from any repo-wide workflow.— blog
- Qwen 3.8 Max is extremely good at knowing what not to do.— reddit ↗
- Qwen3.8-Max is the only thing that comes close to Kimi K3. We need competition in the open-weights space.— hn ↗
- Lately Qwen 3.8 Max has been getting harder to use — even a small request can take half an hour, and it still makes mistakes. It used to run fast; now it's slow and error-prone.— qoder-forum
- It's not that more people are using it — Qwen 3.8 was just released and the official concurrency quota was lowered. Free users queueing is normal; a paid membership lets you skip the queue.— aliyun-ask
- Model wait times are too frequent — it used to be fast, but now basically every task has to wait at least a minute.— aliyun-ask
- Same stupid stuff like qwen3.8-max — degenerates, forgets instructions.— huggingface
- Qwen3.8-Max has been getting harder to use — often a small request takes half an hour to execute and still errors out. It used to be fast; now it's slow and error-prone.— other ↗
“Alibaba's best model yet, and the Chinese + English mix is finally balanced.”
“Open weights plus a huge ecosystem — you can actually build on this.”
“Leaked on Code Arena before launch, so we all knew what was coming.”
FAQ
Is Hunyuan Hy4 or Qwen 3.8 Max more worth it?
Hunyuan Hy4 is more worth it, with a 49% worth-it score vs 43% for Qwen 3.8 Max.
How much does Hunyuan Hy4 cost vs Qwen 3.8 Max?
Hunyuan Hy4: ¥6/¥18 per 1M ($0.83/$2.50) · Qwen 3.8 Max: $2/$6 per 1M tokens.
What do users say about Hunyuan Hy4 vs Qwen 3.8 Max?
Hunyuan Hy4: Horizon Math score is the best among all open models, second only to GPT-5.6 Sol.. Qwen 3.8 Max: Alibaba's best model yet, and the Chinese + English mix is finally balanced..