Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
riknos314
56 days ago
|
parent
|
context
|
favorite
| on:
Kimi-K3 on HuggingFace
Qwen 3.5 27B also scores higher than 3.5 122BA10B. So even in the same generation the smaller dense model outperformed the larger MOE
riknos314
56 days ago
|
next
[–]
https://artificialanalysis.ai/evaluations/artificial-analysi...
apitman
55 days ago
|
prev
[–]
Ah gotcha, I hadn't noticed that. Thanks
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: