Asus Ascent GX10
Ranked 7 of 18 mini PCs by recommenders
Recommended as a compact 128GB CUDA box for local AI development and learning; critics say it is slow for inference and expensive.
Recommended as a compact 128GB CUDA box for local AI development and learning; critics say it is slow for inference and expensive.
46 people gave an opinion
- recommend it
- 21
- criticize it
- 25
- approval share
- 46%
Recommended in 9 of 13 months
- November 2025: recommended
- December 2025: not recommended
- January 2026: recommended
- February 2026: not recommended
- March 2026: not recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 21 people recommend itNeeds 10 or more, met.
- 0.8 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 9 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Good for CUDA and AI development
“I'm in love with my Asus GX10 and have been learning CUDA on it while playing with vllm and such.”
Comment from 13 Apr 2026 in “Taking on CUDA with ROCm: 'One Step After Another'” “if you're only testing things out using CUDA and want 200GB networking and somewhat low power”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “it's exactly as you say. This is a device for AI research and development”
Comment from 1 Jan 2026 in “Dell's version of the DGX Spark fixes pain points” - Comment from 21 May 2026 in “Was my $48K GPU server worth it?”
Large unified memory and fast prefill
“The Spark's GPU gets ~4x the FP16 compute performance of an M3 Ultra GPU on less than half the Mac Studio's total TDP.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “Looks like a pretty useful offering, 128Gb Memory Unified, with the ability to be chained.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “I'd rather have the blazing prefill speed of the Nvidia over my Strix Halo”
Comment from 7 Jul 2026 in “AMD Ryzen AI Halo – $4k AI Dev Kit” - Comment from 10 Nov 2025 in “Asus Ascent GX10”
Better value than rivals
“Buy 4 GX10s with the same money, and enjoy dramatically better models and user scalability.”
Comment from 18 Jun 2026 in “Local Qwen isn't a worse Opus, it's a different tool” “it's the only model in the whole lineup that isn't priced insanely”
Comment from 26 Aug 2026 in “GLM-5.3-Flash” “total cost less than half what he paid, even today”
Comment from 18 Jun 2026 in “Local Qwen isn't a worse Opus, it's a different tool” - Comment from 10 Nov 2025 in “Asus Ascent GX10”
Used as a daily workstation
“if you want desktop consider gx10 given that you need a computer to use this anyway”
Comment from 7 Sep 2026 in “The smallest edge AI device for local LLMs” “I use my ASUS GX10 as my daily driver, my primary workstation.”
Comment from 6 Jul 2026 in “AMD Ryzen AI Halo – $4k AI Dev Kit”
What people warn about
Each mark is one comment
Evasive answers about specifications
“doesn’t seem promising when they ask and then evade a crucial question.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “doesn't seem promising when they ask and then evade a crucial question”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “they're gonna weasel their way around it and not answer properly”
Comment from 10 Nov 2025 in “Asus Ascent GX10” - Comment from 10 Nov 2025 in “Asus Ascent GX10”
- Comment from 10 Nov 2025 in “Asus Ascent GX10”
Slow for LLM inference and training
“it can do impressive stuff but it won't be fast”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “You'd have to be doing something where the unified memory is specifically necessary, and it's okay that it's slow.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “They're not the best choice for anyone who wants to run LLMs as fast and cheap as possible at home.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” - Comment from 11 Nov 2025 in “Asus Ascent GX10”
- Comment from 10 Nov 2025 in “Asus Ascent GX10”
Price seen as high
“The tough part is the price.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “It seamlessly combines Nvidia's price gouging and ASUS's shady tactics.”
Comment from 10 Nov 2025 in “Asus Ascent GX10” “I'm not sure I'd buy an out-there Asus v1 product this expensive.”
Comment from 10 Nov 2025 in “Asus Ascent GX10”
Other mini PCs people recommend
The four most recommended other mini PCs, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Apple Mac mini | 390 | 160 | 71% | 13 | 2 yrs42 owners | Approved |
| NVIDIA DGX Spark | 170 | 121 | 58% | 12 | 6 mo16 owners | — |
| AMD Strix Halo | 130 | 83 | 61% | 13 | 11 mo8 owners | — |
| Framework Desktop | 93 | 26 | 78% | 12 | 3 mo9 owners | Approved |
| Asus Ascent GX10This page | 21 | 25 | 46% | 9 | — | — |
What people said
40 of the 58 comments that mention it, newest first. Each links to the original comment.
“Nvidia's NVFP4 quantisation of Flash Next and offloading the ngram table to NVMe has worked well”
“if you want desktop consider gx10 given that you need a computer to use this anyway”
“I'm constantly running out of room and it being an odd size M.2 slot it's a pain to upgrade”
“it's the only model in the whole lineup that isn't priced insanely”
“I think the Nvidia based Asus is a better choice.”
“I bought the 1T and then did it again.”
“I'd rather have the blazing prefill speed of the Nvidia over my Strix Halo”
“I use my ASUS GX10 as my daily driver, my primary workstation.”
“the Nvidia will be a little faster and a little better supported for inference and training workloads”
“Asus Ascent GX10 is $3999 at various retailers”
“Buy 4 GX10s with the same money, and enjoy dramatically better models and user scalability.”
“total cost less than half what he paid, even today”
“it makes for a great learning platform”
“I'm in love with my Asus GX10 and have been learning CUDA on it while playing with vllm and such.”
“it's exactly as you say. This is a device for AI research and development”
“I'm better off with the DG10 than AMD or Apple's systems.”
“Running LLMs will be slow and training them is basically out of the question.”
“Now we can see why they avoided giving a straight answer.”
“it can do impressive stuff but it won't be fast”
“doesn't seem promising when they ask and then evade a crucial question”
“The Asus clustering speed is not limited to p2p.”
“The Spark's GPU gets ~4x the FP16 compute performance of an M3 Ultra GPU on less than half the Mac Studio's total TDP.”
“You'd have to be doing something where the unified memory is specifically necessary, and it's okay that it's slow.”
“The GX10 will likely be quite disappointing in terms of tokens per second”
“Taiwanese companies are legendary for producing baller hardware with terrible marketing and documentation that answers important questions.”
“Overall it feels close to full upstream support”
“It's very, very good as an ARM Linux development machine”
“The tough part is the price.”
“Laptop-class bandwidth without that annoying portability.”
“I'm frankly happy to see workstation-class AArch64 hardware available to regular consumers.”
“They're not the best choice for anyone who wants to run LLMs as fast and cheap as possible at home.”
“In this case the Asus "tax" is negative $1,000.”
“if you're only testing things out using CUDA and want 200GB networking and somewhat low power”
“It seamlessly combines Nvidia's price gouging and ASUS's shady tactics.”
“Simon Willison seems to like his”
“fails to comprehend the question: ignoring the word”
“they're gonna weasel their way around it and not answer properly”
“doesn’t seem promising when they ask and then evade a crucial question.”
“I'm not sure I'd buy an out-there Asus v1 product this expensive.”
“Looks like a pretty useful offering, 128Gb Memory Unified, with the ability to be chained.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.