Intel Panther Lake Core Ultra X9 388H performance analysis - Outpaces Arrow Lake and exceeds Zen 5 in efficiency

[Worgarthe]

Lmao they'll defend AMD's neverending rebranding to the death too but heaven forbid Intel or nVidia actually release something good for once...

My favourite is:

• When AMD is beating Intel in benchmark ABC then it's always "Hahah, Intel is so trash, AMD is beating it in the ABC"
• When Intel is beating AMD in that same benchmark: "The ABC is a complete irrelevant trash, AMD is better."

Perfect recent example from this article: Intel Core Ultra X9 388H chip crushes AMD Ryzen AI 9 HX 470, Core Ultra 9 285H in latest Passmark appearance

Passmark is a worthless benchmark.

Intel's Core Ultra X9 388H is definitely promising

Unfortunately, I think it'll be short-lived now that N1X is going to be announced officially in a few months. Proper Nvidia drivers with RTX 5070 performance in an APU. Why would anyone get this anymore?

AMD fanatics are truly something else...

[M2026]

@Consumers_Are_Citizens
,,It's a game changer in heat, noise and battery life."
Well, what do you think is the difference between TDP 28W vs 28W (Intel vs AMD)?
What do you think is the difference between TDP 45W vs 45W (Intel vs AMD)?

In the case of Intel, 28/45W is less than 28/45W in the case of AMD? You can't be serious! :)

In the case of passmark the difference is "huge" - less than 4%, in the case of cinebench less than 3%. Is this the mentioned "gamechanger"?

Moreover, everyone is now praising iGPU on Intel, but professional software needs a powerful GPU, not this...!
But okay, in the case of the praised iGPU I will use slightly modified words of "kninez":
There is no "Intel's iGPU", there is only TSMC's iGPU glued to Intel. loool

There are not enough tests yet, let's wait for reviews of really powerful workstations. But if these numbers are confirmed, it will be just another chapter in Intel's "How to rip off the customer" manual...I don't want to be a bad prophet, but compare the presentation of Intel, Nvidia and AMD at CES 2026 - that says something.

[Worgarthe]

There is no "Intel's iGPU", there is only TSMC's iGPU glued to Intel. loool

So you just described AMD, Apple and Nvidia. Nothing but TSMC's dGPUs and iGPUs. Coping level pro.

[M2026]

"So you just described AMD, Apple and Nvidia. Nothing but TSMC's dGPUs and iGPUs. Coping level pro."
Maybe you should take it in context, but you would have to read that with understanding.

[Worgarthe]

"So you just described AMD, Apple and Nvidia. Nothing but TSMC's dGPUs and iGPUs. Coping level pro."
Maybe you should take it in context, but you would have to read that with understanding.

Yes, the context you mentioned it quite literally - it's not Intel's iGPU but TSMC's. And, of course, glorified SCAMD:

Moreover, everyone is now praising iGPU on Intel, but professional software needs a powerful GPU, not this...!
But okay, in the case of the praised iGPU I will use slightly modified words of "kninez":
There is no "Intel's iGPU", there is only TSMC's iGPU glued to Intel. loool

There are not enough tests yet, let's wait for reviews of really powerful workstations. But if these numbers are confirmed, it will be just another chapter in Intel's "How to rip off the customer" manual...I don't want to be a bad prophet, but compare the presentation of Intel, Nvidia and AMD at CES 2026 - that says something.

Sure, not understanding yourself is your starting point, but that's a whole different story...

[Citizen_not_Consumer]

Intel's 1.8nm appears to be a total failure:

1. All potential customers (Apple, Qualcomm, etc) who tested Intel's 1.8nm left running to queue up at TSMC and Samsung despite the increased prices and long waiting times. Intel intself is using TSMC for critical components (iGPU, and many CPU tiles still).
2. Intel's Panther Lake 1.8nm CPU efficiency is equal to 4nm TSMC's.
3. In few months Zen 6 is released (already in production for EPYC) and Intel will be trailing again for years.

The only thing Intel is good at, is spending taxpayers money by receiving huge subsidies (just received another $150 billion from Pentagon) instead of these money going to better education or health care so future generations can be better educated and compete internationally!

[opckieran]

@Tyler
1. yes, it is faster – 2,6%, that is so interesting or even a "gamechanger"? 1,8 nm vs 4 nm and less, than 3%???
2. I am NOT an AMD fan – the opposite, that is why this is a disappointment for me as well, but let´s wait for the workstations
3. the iGPU is great, but only for MS Office, accounters etc. Not a single CAD/CAM guy will buy the Panther Lake laptop because of the iGPU!

@opckieran
,,Time for the fanboys to wake up..."
Yes, it is about the time to wake up – FOR YOU! Loool
Hype for how long, more than 1 year and then 3% better (sorry 2,6% :), than the 370 = 1,8nm vs 4nm? loool

Keep coping! You honestly believe customers make their final purchasing decision based on the *manufacturing node* as opposed to performance and efficiency?? And the GPU is 70% faster than 370's, not 3% (CPU perf parity doesn't help AMD either btw). Nobody said this iGPU was aimed at CAD/CAM. That's you shifting the goalposts.

Intel's 1.8nm appears to be a total failure:

1. All potential customers (Apple, Qualcomm, etc) who tested Intel's 1.8nm left running to queue up at TSMC and Samsung despite the increased prices and long waiting times. Intel intself is using TSMC for critical components (iGPU, and many CPU tiles still).
2. Intel's Panther Lake 1.8nm CPU efficiency is equal to 4nm TSMC's.
3. In few months Zen 6 is released (already in production for EPYC) and Intel will be trailing again for years.

The only thing Intel is good at, is spending taxpayers money by receiving huge subsidies (just received another $150 billion from Pentagon) instead of these money going to better education or health care so future generations can be better educated and compete internationally!

No actual consumers give a sh*t about manufacturing nodes... It's just e-peen measuring for nerds. Real customers care about the final product.

Zen 6 better be good for AMD's sake. After Panther Lake, Zen 6 needs to be +20% over Zen 5 if they want to stay competitive in the laptop space!

And at this point, I'd rather subsidize companies than get any more Learing Centers! Putting our education in the hands of foreigners... All that accomplished was us getting scammed to high heaven! I didn't vote for this woke-mandated nonsense!

[Worthless]

AMD fanatics are truly something else...

There are quite a few in this thread but not me lol. My comment was said solely out of passionate hatred for Passmark in general. Wasn't even thinking of AMD at all.

[Consumers_Are_Citizens]

M2026, it's really not that deep. Every single reviewer who has tested this chip says they were impressed with noise and thermals compared to AMD. From here we can infer it's a noticeable step up in efficiency without even diving into the numbers.

But if you insist on stats and graphs, I already stated earlier ThePhawx made a 1 hour video testing panther lake vs strix halo, at low tdps (15W and 30W). At 15W in the majority of cases it's beating strix halo. At 30W is very similar or almost as fast as strix halo.

Mind you, Phawx has previously been more of an AMD fan in recent years and was previously not enthusiastic at all for Intel's panther lake when asked. He is literally the definitive authority when it comes to amd handheld tdp efficiency testing and even he has stated, that intel is back now.

[Worgarthe]

But if you insist on stats and graphs, I already stated earlier ThePhawx made a 1 hour video testing panther lake vs strix halo, at low tdps (15W and 30W). At 15W in the majority of cases it's beating strix halo. At 30W is very similar or almost as fast as strix halo.

Great comment, and the video you're talking about is also quite excellent. However, one small correction regarding the part in bold - it is beating it in all cases as the Strix Halo cannot even run at less than 13W, and it's not even close the Panther Lake up until 20W. Everything else you said is correct 👍