News:

Willkommen im Notebookcheck.com Forum! Hier können Sie über alle unsere Artikel und allgemein über notebookrelevante Dinge diskutieren. Viel Spass!

Main Menu

Asus ZenBook S14 2026 Laptop Review - Brighter OLED, but a slow Panther Lake iGPU

Started by Redaktion, Today at 02:13:55

Previous topic - Next topic

Redaktion

Asus refreshed its high-end Zenbook S14 with a new Panther Lake processor from Intel, a slightly bigger battery as well as a brighter OLED screen. While the CPU performance has increased, the GPU performance is lower than before.

https://www.notebookcheck.net/Asus-ZenBook-S14-2026-Laptop-Review-Brighter-OLED-but-a-slow-Panther-Lake-iGPU.1342874.0.html


And on top of that

Quote from: 123 on Today at 03:02:07y'all lost your mind with those prices
And on top of that:
32 GB RAM is enough for about 16,000 to 32,000 tokens of context for the current SOTA, for its size, model, UD-Q4_K_XL (22.9 GB) quant: huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:
32 GB RAM - 8 GB for the OS itself = 24 GB free for fitting an AI LLM model.
24 GB - 22.9 = ~1 GB. 1 GB is enough for about 16,000 context tokens[1].
Another confirmation:
Quote from: reddit.com/r/LocalLLaMA/comments/1sq94qx/is_anyone_getting_real_coding_work_done_with.. I've come to the conclusion that (1) 32768 is the biggest context I can get away with in an adequately smart model, and (2) it just ain't enough.
Not only agentic workloads often require over 100,000 of tokens. An 8 GB VRAM GPU (but then it would not be as thin an light and would basically be a gaming laptop) or 48 GB RAM config would solve this issue.

[1] reddit.com/r/LocalLLaMA/comments/1tvluaj/how_much_vram_needed_for_qwen_36_27b_q8_with_262k

And on top of that

Forgot to say: A UD-Q4_K_XL quant is known as a best for the bang quant. Good for about 80k to 120k tokens of context.

Quick Reply

Name:
Email:
Verification:
Please leave this box empty:
Shortcuts: ALT+S post or ALT+P preview