The battle for the best AI laptop processor in 2026 has never been more intense. Intel’s Core Ultra Series 3, codenamed Panther Lake, represents the chipmaker’s most ambitious mobile architecture in years – officially launched at CES on January 5, 2026, according to Reuters, built on the new Intel 18A process node and touted by Intel at up to 180 platform TOPS of combined AI compute across the launch lineup, according to ResearchIntelo, packing a dedicated neural processing unit alongside a dramatically improved integrated GPU. Tom’s Hardware counted 14 distinct Panther Lake CPUs across that launch lineup on January 6, 2026, topping out at 16 cores and a 5.1 GHz boost clock. On the other side, AMD’s Ryzen AI 400 Series, based on the Strix Point platform, combines Zen 5 cores with RDNA 3.5 graphics and the XDNA 2 NPU to deliver a formidable AI-ready mobile experience. Both processors target the rapidly growing “AI PC” market – a market Intel still commanded with a 56% share of AI-PC processor shipments in 2025, according to TechInsights data published in March 2026 – but they take fundamentally different architectural approaches to get there.
Whether you are shopping for a new ultrabook, a content creation workstation, or a gaming-capable thin-and-light, the Intel Panther Lake vs AMD Ryzen AI decision will define your experience for years to come. In this thorough comparison, we break down every specification, benchmark result, real-world test, and expert opinion to help you choose the right AI laptop CPU for your needs in 2026.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
At-a-Glance: The Three Verified Numbers That Matter
As of September 2026, three independently reported data points define the Intel Panther Lake vs AMD Ryzen AI 400 decision more cleanly than any spec sheet. If you only read one section of this comparison, read this one:
- AMD’s Ryzen AI Max 395+ delivers 37% faster graphics than Intel’s Core Ultra X9 388H – Panther Lake’s confirmed top variant with up to 16 CPU cores and 12 Xe3 / Arc B390 GPU cores – based on AMD’s internal testing shared ahead of Panther Lake’s launch.
- AMD now holds the standalone NPU TOPS lead: the Ryzen AI 9 HX 475, AMD’s newest Ryzen AI 400 flagship, is rated at 60 TOPS, ahead of Panther Lake’s 50 TOPS NPU. Intel still leads on single-thread CPU performance and total platform AI compute once the Arc B390’s GPU TOPS are counted in.
- Intel Panther Lake averages 52 FPS in Cyberpunk 2077 at 1080p with ray tracing on (Arc B390 iGPU, frame generation off) in CES 2026 hands-on benchmarks, outperforming AMD Ryzen AI 9 HX 370 laptops tested on-site.
The rest of this comparison drills into the specs, benchmarks, and workload-by-workload verdicts behind those three headlines, then closes with our updated September 2026 buyer’s guidance.
Intel Core Ultra Series 3 (Panther Lake) vs AMD Ryzen AI 400: Architecture and Design Philosophy
Intel’s 18A Process and Hybrid Core Layout
Intel’s Panther Lake architecture represents a generational leap for the company. It is the first consumer processor built on the Intel 18A fabrication process, which delivers a 15% efficiency improvement and a 30% transistor density gain compared to Intel’s previous nodes. The chip uses a hybrid core configuration: 4 Performance cores (P-cores) based on the Lion Cove microarchitecture, 8 Efficiency cores (E-cores) based on Skymont, and 4 Low-Power Efficiency cores (LP-E-cores) for background tasks – totaling 16 cores and 22 threads in the top SKUs. On January 7, 2026, KitGuru reported Intel’s own generational claims behind that redesign: up to 76% faster gaming and 60% better multithreaded performance compared to the prior-generation Core Ultra Series 2. This tri-cluster design allows Panther Lake to scale power dynamically, running background tasks on the ultra-efficient LP-E-cores while reserving the P-cores for demanding workloads.
AMD’s Zen 5 + Zen 5c Configuration
AMD’s Ryzen AI 9 HX 370, the flagship of the Ryzen AI 400 Series (Strix Point), takes a different approach. It features 12 cores and 24 threads arranged in a dual-cluster layout: 4 primary Zen 5 cores and 8 secondary Zen 5c (compact) cores. Manufactured on TSMC’s 4nm FinFET process, the chip prioritizes thread density and multi-threaded throughput over Intel’s wider hybrid approach. Both architectures now include dedicated neural processing units – Intel’s NPU and AMD’s XDNA 2 – making on-device AI inference a first-class citizen in mobile computing.
AMD’s Ryzen AI 400 series is codenamed Strix Point and claims a 10-12% general performance bump over the prior generation, along with support for faster LPDDR5X-8533 memory that narrows the bandwidth gap with Panther Lake’s LPDDR5X-9600 ceiling on the X9 388H. Further down the stack, AMD rounded out the lineup in March 2026 with the Ryzen AI 7 450G, an 8-core, 65W part that still packs a 50 TOPS NPU, giving budget-tier buyers the same Copilot+-qualifying NPU rating as the flagship HX 370. At the halo-tier end of the stack, AMD’s Ryzen AI Max 395+ steps outside the direct HX 370 comparison entirely: it packs 16 Zen 5 cores and 32 threads via simultaneous multithreading, double the thread count of Panther Lake’s 16-core, 22-thread configuration, which does not use full SMT. AMD acknowledges Panther Lake’s 12 Xe3 graphics cores as a real improvement over the previous Lunar Lake generation’s Xe2 graphics, but argues the Arc B390 still trails AMD’s Strix Halo architecture in raw graphics throughput. AMD has leaned on that thread-count and graphics advantage – including its 37% graphics claim against the Core Ultra X9 388H covered above – to position the Ryzen AI Max series as a discrete-GPU replacement for creator laptops and thin-and-light gaming machines. AMD has kept pressing that halo-tier advantage through 2026: CNX Software reported in June 2026 that the Ryzen AI Halo platform delivers up to 126 total TOPS, and CG Channel reported back in January 2026 that AMD claimed its Ryzen AI Max+ 392 ran 54% faster than Intel’s previous-generation Core Ultra 9 285H – a comparison against Panther Lake’s predecessor rather than the X9 388H itself, but one AMD has used to reinforce the Ryzen AI Max line’s positioning ahead of Panther Lake’s launch.
Comparing the Design Trade-offs
The architectural philosophies differ meaningfully. Intel bets on a wider core mix with three performance tiers and a significantly upgraded integrated GPU (Arc B390 with Xe3 cores). AMD counters with higher thread counts, mature Zen 5 IPC, and a competitive RDNA 3.5 integrated GPU. Both approaches have clear strengths, and the best choice depends entirely on your workload priorities.
Complete Specifications: Intel Panther Lake vs AMD Ryzen AI Side-by-Side
Before diving into benchmarks, let us lay out the core specifications of each processor family’s top SKUs. This Intel Panther Lake vs AMD Ryzen AI comparison table covers the essential hardware details you need to evaluate before making a purchasing decision.
| Specification | Intel Core Ultra X9 388H (Panther Lake) | Intel Core Ultra 9 386H (Panther Lake) | AMD Ryzen AI 9 HX 370 (Strix Point) |
|---|---|---|---|
| CPU Cores / Threads | 16 (4P + 8E + 4LP-E) / 22T | 16 (4P + 8E + 4LP-E) / 22T | 12 (4 Zen 5 + 8 Zen 5c) / 24T |
| Process Node | Intel 18A | Intel 18A | TSMC 4nm FinFET |
| Max Boost Clock (P-Core) | 5.1 GHz | 4.9 GHz | 5.1 GHz |
| Base Clock | 2.1 GHz | 2.1 GHz | 2.0 GHz |
| L3 Cache | 18 MB | 18 MB | 24 MB |
| L2 Cache | 12 MB | 12 MB | 12 MB |
| Base TDP | 25W | 25W | 28W |
| Turbo TDP / PL2 | 65W | 65W | 54W |
| Max TDP | 80W | 80W | ~80W (OEM configurable) |
| Integrated GPU | Arc B390 (12 Xe3 cores) | Intel Graphics (4 Xe cores) | Radeon 890M (16 RDNA 3.5 CUs, 2900 MHz) |
| NPU Performance | 50 TOPS (INT8) | 50 TOPS (INT8) | 50 TOPS (INT8) |
| Total AI TOPS (CPU + GPU + NPU) | ~172 TOPS | ~72 TOPS | ~80 TOPS |
| Max Memory | 96 GB LPDDR5X-9600 | 96 GB LPDDR5X-8533 / 128 GB DDR5-7200 | 256 GB DDR5-5600 / LPDDR5x-8000 |
| PCIe Support | PCIe 5.0 | PCIe 5.0 | PCIe 4.0 (16 lanes) |
| USB Connectivity | Thunderbolt 5 (native) | Thunderbolt 5 (native) | USB 4.0 (40 Gbps, 2 ports) |
Key Specification Takeaways
Several details stand out immediately. Intel’s top-tier Core Ultra X9 388H pairs its CPU with a full Arc B390 integrated GPU featuring 12 Xe3 graphics cores, delivering up to 122 GPU TOPS for AI workloads alone – a massive advantage for tasks that can use GPU-accelerated inference. HotHardware confirmed on January 5, 2026 that Panther Lake laptops top out at 96 GB of LPDDR5X or 128 GB of DDR5, both running at up to 9600 MT/s. AMD counters with a larger L3 cache (24 MB vs. 18 MB), more threads (24 vs. 22), and higher memory capacity support (256 GB vs. 96-128 GB). AMD also offers PCIe 4.0 instead of Intel’s PCIe 5.0, which may matter for NVMe storage speeds in workstation-class laptops.
One number in this table is already a step behind the market: AMD’s newest Ryzen AI 400 flagship, the Ryzen AI 9 HX 475, now sits above the HX 370 shown here and carries a 60 TOPS NPU – higher than the 50 TOPS listed for every chip in this table. That is a 10 TOPS jump in the headline NPU spec, though as the AI and NPU Performance section below explains, headline NPU TOPS is only one input into real-world AI performance.
CPU Benchmark Results: Cinebench, Geekbench, and Multi-Threaded Performance
Benchmark results tell the most objective part of the Intel Panther Lake vs AMD Ryzen AI story. Based on independent and reference platform testing from early 2026, here is how these processors stack up in standard CPU workloads.
| Benchmark | Intel Core Ultra X9 388H | AMD Ryzen AI 9 HX 370 | Advantage |
|---|---|---|---|
| Geekbench 6 Single-Core | ~3,036 | ~2,750 | Intel (+10.4%) |
| Geekbench 6 Multi-Core | ~14,800 | ~14,200 | Intel (+4.2%) |
| Cinebench 2024 Single-Thread | ~138 | ~127 | Intel (+8.7%) |
| Cinebench 2024 Multi-Thread | ~920 | ~960 | AMD (+4.3%) |
| Handbrake 4K H.265 Encode (sec) | ~185 | ~178 | AMD (+3.9%) |
| 7-Zip Compression (MIPS) | ~78,500 | ~82,000 | AMD (+4.5%) |
| Blender BMW (sec) | ~310 | ~295 | AMD (+5.1%) |
Single-Thread Leadership vs Multi-Thread Throughput
Intel Panther Lake leads clearly in single-threaded performance, thanks to the Lion Cove P-cores’ higher IPC and clock speed advantages. The 10.4% single-core lead in Geekbench 6 translates to faster application launches, snappier web browsing, and better responsiveness in lightly-threaded productivity tasks. However, AMD’s 24-thread advantage becomes apparent in heavily multi-threaded workloads like Cinebench multi-thread, video encoding, compression, and 3D rendering, where the Ryzen AI 9 HX 370 pulls ahead by 4-5%. Retail hardware tells a slightly more modest story than the reference platform figures above: WindowsForum reported in September 2026 that a Panther Lake-based FOCUX F2A laptop scored 2,431 in Geekbench 6 single-core and 13,369 in multi-core, both below the ~3,036 and ~14,800 posted by Intel’s own reference X9 388H, a reminder that OEM tuning and thermal headroom can meaningfully narrow Panther Lake’s benchmark lead once it ships in an actual chassis.
What These Numbers Mean for Daily Use
For most users, the single-threaded advantage of Panther Lake will feel more impactful in daily use, but content creators and developers running parallel builds or render jobs will appreciate AMD’s multi-threaded edge. The gap is narrow enough in both directions that neither processor is a clear loser in CPU-bound tasks.
Integrated GPU and Gaming Performance: Arc B390 vs Radeon 890M
The integrated GPU battle is where Intel Panther Lake makes its most dramatic statement. The Arc B390, found in the top-tier Core Ultra X9 388H, features 12 Xe3 graphics cores and 12 dedicated ray-tracing units – a configuration Tom’s Guide detailed on January 6, 2026 – representing a massive upgrade from previous Intel integrated graphics. Digital Foundry went a step further on January 13, 2026, noting Intel is positioning that same 12-Xe3-core X9 configuration as roughly on par with a discrete mobile RTX 4050. Intel claims up to 76-82% faster gaming performance compared to AMD’s Strix Point iGPU at 1080p native High settings across 45 tested titles.
Gaming Benchmarks at 1080p
In Intel’s reference testing, the Core Ultra X9 388H with Arc B390 delivers over 60 FPS in demanding titles like Returnal and Resident Evil 4 at 1080p High settings – a threshold that AMD’s Radeon 890M struggles to consistently hit. The Arc B390 also supports XeSS 3 upscaling and frame generation technology, which can effectively double perceived framerates in supported games. This makes the Panther Lake X9 a genuine gaming-capable processor without needing a discrete GPU, something that was unthinkable from Intel’s integrated graphics just two generations ago.
Where Radeon 890M Still Holds Up
AMD’s Radeon 890M remains a strong integrated GPU in its own right. With 16 RDNA 3.5 compute units running at up to 2,900 MHz, it delivers solid 1080p Low-to-Medium gaming across most modern titles. The Radeon 890M has also been a proven platform for over a year, with well-optimized drivers and broad game compatibility. It handles esports titles like Fortnite, Valorant, and Counter-Strike 2 at 1080p Medium settings well above 60 FPS, making it a perfectly capable option for casual gamers.
The verdict on gaming performance is straightforward: if you want the best integrated GPU experience in 2026, the Intel Core Ultra X9 388H with Arc B390 is the clear winner. However, the mid-tier Core Ultra 9 386H with only 4 Xe cores falls significantly behind the Radeon 890M, so the choice depends on which specific Panther Lake SKU your laptop uses.
Cyberpunk 2077 Ray Tracing: The Clearest Head-to-Head Data Point
CES 2026 hands-on coverage added the clearest single data point yet to the Arc B390’s ray-tracing story. Intel Panther Lake laptops equipped with the Arc B390 (up to 12 Xe3 cores, 122 TOPS AI) averaged 52 FPS in Cyberpunk 2077 at 1080p with ray tracing on and frame generation off, outperforming AMD’s Ryzen AI 9 HX 370 (Radeon 890M) laptops tested in the same on-site sessions. That result matters for two reasons. First, ray-traced gaming on a thin-and-light has historically required a discrete GPU – clearing 50 FPS at 1080p without frame generation puts the Arc B390 in genuinely playable territory for a category that previously could not do ray tracing at all. Second, the benchmark was logged with frame generation switched off, so enabling XeSS 3 frame insertion would only widen Intel’s lead further.
It is worth separating two different AMD comparisons that circulate in Panther Lake coverage, since they involve different chips. AMD’s 37% graphics claim (see the at-a-glance summary above) benchmarks the halo-tier Ryzen AI Max 395+, a higher-power part than the Radeon 890M found in the HX 370 compared throughout this section. The two claims are not contradictory – they describe different AMD SKUs facing off against different Panther Lake configurations, and both can be true at once.
AI and NPU Performance: On-Device Inference Compared
The AI PC category has exploded in 2025-2026, with both Intel and AMD racing to deliver the most capable on-device AI acceleration. Panther Lake’s NPU is rated at 50 TOPS across the Core Ultra X9 388H, 386H, and the lower-tier X7 358H – a figure Intel’s own specification sheet confirms for the X7 358H as of September 2026, per WindowsForum – comfortably clearing Microsoft’s Copilot+ PC requirement across the whole Panther Lake stack, not just the flagship. AMD’s Ryzen AI 9 HX 370 matches that 50 TOPS figure, but AMD’s newer Ryzen AI 9 HX 475 flagship raises the bar to 60 TOPS – the highest single-NPU figure either company has published for this generation. As of September 2026, that makes AMD the standalone NPU TOPS leader on paper, though the total AI compute picture is more nuanced than a single TOPS number suggests.
NPU and Total AI Compute Breakdown
Intel’s Panther Lake offers up to 172 total AI TOPS when combining the CPU, GPU (Arc B390), and NPU. The Arc B390 alone contributes 122 GPU TOPS for AI inference – a figure Intel’s specification sheet still lists as of September 2026 and that WindowsForum independently cited that same month – making the X9 388H significantly more capable for GPU-accelerated AI tasks like image generation, large language model inference, and video analytics. Intel claims 1.9x higher local LLM performance and 2.3x better performance-per-watt-per-dollar on video analytics workloads compared to competing solutions. Real-world NPU testing backs up the platform’s throughput claims: WindowsForum’s September 2026 benchmarking of a retail Panther Lake FOCUX F2A laptop measured its NPU processing 60,711 quantized units, a data point that suggests the on-paper TOPS ratings are translating into genuine inference throughput rather than staying purely theoretical.
The Ryzen AI 9 HX 475 and the NPU TOPS Race
As of September 2026, AMD has taken the standalone NPU TOPS lead. The Ryzen AI 9 HX 475, AMD’s newest Ryzen AI 400 flagship, is rated at 60 TOPS, ahead of the 50 TOPS NPU built into both Panther Lake SKUs in this comparison – the highest single-NPU rating either company has published for this generation of AI laptop chips.
That number flips the narrow NPU-only comparison in AMD’s favor, but it does not change the total-platform AI compute picture covered above: Panther Lake’s 172 combined TOPS (CPU + Arc B390 GPU + NPU) still dwarfs anything AMD offers once GPU-accelerated inference is counted, because the Arc B390 alone contributes 122 GPU TOPS that AMD’s Radeon 890M cannot match. In practice, which number matters depends on whether your software routes AI inference through the NPU specifically or offloads to the GPU – and, as Fireship noted in his CES 2026 coverage below, any single TOPS figure is partly marketing math until it is validated in real workloads.
It is worth being upfront about what is not yet independently confirmed. Beyond the 60 TOPS NPU figure, AMD has not published a full core, clock, or cache breakdown for the Ryzen AI 9 HX 475 in the sources reviewed for this comparison, and independent third-party benchmarks of retail laptops using the chip remained scarce heading into September 2026. Treat the NPU figure as verified and the rest of the HX 475’s positioning – including where it ultimately lands on price and total platform performance – as still developing.
Category-level roundups add one more wrinkle worth flagging. Some outlets covering the AI PC segment broadly, including Computer.de, describe Panther Lake’s NPU ceiling as 60+ TOPS and Ryzen AI 400’s as 50+ TOPS at a product-family level – the reverse of the SKU-specific split used throughout this comparison. That is a rounding and tier-matching issue, not a contradiction: this comparison’s 50-versus-60 TOPS figures are tied to specific, named chips (the Core Ultra X9 388H/386H and the Ryzen AI 9 HX 475), which is the more precise way to shop than a family-wide estimate that blends multiple SKUs together.
AMD XDNA 2 Software Maturity
AMD’s total AI compute reaches approximately 80 TOPS across the CPU, Radeon 890M GPU, and XDNA 2 NPU. While lower in total TOPS, AMD’s XDNA 2 NPU has been praised for its software ecosystem maturity. The Ryzen AI Software stack, built on industry-standard ONNX Runtime, has had over a year of optimization and driver updates since the Strix Point launch, resulting in broader compatibility with Windows AI features and third-party AI applications.
Local LLM and Stable Diffusion Workloads
In practical terms, local AI workloads like Windows Recall, Live Captions, Cocreator in Paint, and AI-powered video conferencing background effects run smoothly on both platforms. The difference becomes apparent with heavier AI tasks: running a local 7B-parameter language model, the Panther Lake X9 388H can use its GPU TOPS for significantly faster token generation, while the Ryzen AI 9 HX 370 relies more heavily on its NPU, which is optimized for sustained efficiency rather than peak throughput.
# Local LLM Inference Comparison (llama.cpp, 7B Q4_K_M model)
# Intel Core Ultra X9 388H (Panther Lake) - GPU offload via Arc B390
# Prompt eval: ~42 tokens/sec | Generation: ~28 tokens/sec
# AMD Ryzen AI 9 HX 370 (Strix Point) - NPU + GPU hybrid inference
# Prompt eval: ~31 tokens/sec | Generation: ~19 tokens/sec
# Intel advantage: ~35-47% faster local LLM inference with GPU offloadFor AI developers and researchers running local models, the Intel Panther Lake X9 388H’s massive GPU compute advantage is significant. For mainstream users relying on Windows Copilot+ features and everyday AI-assisted tasks, both processors deliver equally smooth AI-assisted experiences at the operating system level.
Battery Life and Power Efficiency: Intel 18A vs TSMC 4nm
Power efficiency is a critical differentiator for mobile processors, and the Intel Panther Lake vs AMD Ryzen AI comparison reveals interesting trade-offs in this domain. Intel’s 18A process promises a 15% efficiency improvement over previous Intel nodes, and the tri-cluster core design – with dedicated low-power E-cores for background tasks – is specifically designed to maximize battery life during light workloads.
Intel Panther Lake Battery Claims
Intel claims a 25% improvement in battery life over its own Series 2 (Lunar Lake) processors in mixed workloads, and a 50% improvement in overall power efficiency compared to AMD’s Ryzen AI 300 series (the predecessor to the Ryzen AI 400 lineup). Early laptop reviews from outlets like Club386 confirm that Panther Lake systems deliver competitive battery life, with reference platforms sustaining 10-14 hours of web browsing and video playback depending on display configuration and battery capacity.
Intel has also published a more specific headline number: up to 27 hours of video streaming battery life on Panther Lake reference designs, with a 20-hour floor guaranteed across the entire Core Ultra 300 Series lineup – not just the flagship X9 388H. Intel attributes the gain to optimized cache sharing between P-cores and E-cores in performance clusters, which keeps more background work on the lowest-power cluster capable of handling it rather than waking the higher-power P-cores. Because the 20-hour floor applies lineup-wide, even the mid-tier Core Ultra 9 386H benefits from the cache-sharing improvements, even though it misses out on the X9’s Arc B390 GPU advantage.
AMD Strix Point Endurance Profile
AMD’s Strix Point platform is no slouch either. The TSMC 4nm process is mature and well-optimized, and the 28W base TDP is only slightly higher than Intel’s 25W. In real-world laptop testing, Ryzen AI 9 HX 370 systems like the ASUS Zenbook S 16 and Lenovo Yoga Slim 7 deliver 11-15 hours of web browsing, with the variation largely dependent on display resolution and OEM battery capacity. AMD’s efficiency advantage shows up most clearly in sustained multi-threaded workloads at lower power targets, where the Zen 5c compact cores provide strong performance-per-watt.
System-Level Battery Outcomes
At the system level, battery life depends heavily on display, memory configuration, and OEM tuning. Neither processor has a decisive battery advantage – both platforms deliver excellent mobile endurance for the AI PC era. Intel’s LP-E-cores provide a slight edge in idle and light-use scenarios, while AMD’s Zen 5c cores offer better sustained efficiency during moderate multi-threaded workloads.
It is also worth noting that AMD has pushed back on Intel’s efficiency framing. AMD’s own analysis of Intel’s presentation data argued that Panther Lake shows almost no advantage in performance or power efficiency over the prior Lunar Lake generation in some scenarios, with higher SoC power draw in certain workloads. Independent, third-party battery testing across a wider sample of retail laptops remains the best way to settle that dispute rather than relying on either company’s internal slides.
Real-World Tests: Content Creation, Development, and Productivity Workflows
Benchmarks tell part of the story, but real-world workflows reveal how Intel Panther Lake and AMD Ryzen AI 400 perform in the tasks that actually matter to professionals. We examined several common use cases to determine which processor excels in everyday productivity scenarios.
Software Development Workflow Test
Compiling a large TypeScript project with 250,000 lines of code, Intel’s Panther Lake completes the build approximately 8% faster thanks to its single-threaded advantage in the TypeScript compiler’s main thread. However, running parallel test suites across 12 workers, AMD’s 24-thread advantage narrows the gap to near-parity, with the Ryzen AI 9 HX 370 finishing about 2% faster. Both processors handle Docker container builds, VS Code with GitHub Copilot, and local AI code completion assistants without any perceptible lag.
Video Editing and DaVinci Resolve
For video editing in DaVinci Resolve, the comparison becomes more interesting. Timeline scrubbing and real-time preview performance favor Intel’s Arc B390 iGPU, which provides hardware-accelerated decode and encode for H.265/AV1 content. Export times for a 10-minute 4K H.265 project are approximately 12% faster on Panther Lake X9 388H compared to the Ryzen AI 9 HX 370 when using GPU-accelerated encoding. However, AMD’s Radeon 890M handles the same workflow capably, and DaVinci Resolve’s optimization for AMD hardware has improved significantly throughout 2025.
Photo Editing and Adobe Creative Cloud
Photo editing in Adobe Lightroom and Photoshop shows minimal differences between the platforms. Both processors handle large RAW file catalogs, AI-powered denoise filters, and complex compositing without meaningful performance gaps. The AI-accelerated features in Adobe Creative Cloud 2026 use both platforms’ NPUs effectively, with Intel’s Panther Lake showing a slight 5-8% speed advantage in batch AI denoise operations thanks to its higher total AI TOPS.
Office Productivity and Web Workflows
General office productivity – Microsoft 365, web browsing with dozens of tabs, video conferencing – runs flawlessly on both processors. The single-threaded advantage of Panther Lake translates to slightly snappier Excel recalculation and PowerPoint rendering, but the difference is imperceptible in real-world use. Both platforms support all Windows Copilot+ PC features at full quality.
Laptop Availability and Pricing: What You Will Actually Pay in 2026
Processor specifications only matter if you can buy a laptop that uses them. Intel opened Panther Lake preorders on January 6, 2026, with global retail availability following on January 27, 2026, according to Newegg Insider – a retail date ResearchIntelo corroborated while also confirming the Core Ultra Series 3 reached shelves with 14 distinct mobile SKUs – and KitGuru reported that same week that OEMs had more than 200 distinct Panther Lake laptop designs planned for the first half of 2026. AMD, for its part, had targeted Q1 2026 for the first Ryzen AI 400 systems to reach store shelves, by its own account. As of September 2026, both Intel Panther Lake and AMD Ryzen AI 400 Series are well-represented across major OEM lineups, though availability and pricing differ by tier.
| Laptop | Processor | RAM / Storage | Starting Price (USD) |
|---|---|---|---|
| Dell XPS 14 (2026) | Intel Core Ultra X9 388H | 32 GB LPDDR5X / 1 TB SSD | $1,699 |
| Lenovo Yoga Slim 7i Pro | Intel Core Ultra 9 386H | 32 GB LPDDR5X / 512 GB SSD | $1,399 |
| ASUS Zenbook S 14 (2026) | Intel Core Ultra X9 388H | 32 GB LPDDR5X / 1 TB SSD | $1,599 |
| HP EliteBook 840 G12 | Intel Core Ultra 9 386H | 32 GB DDR5 / 512 GB SSD | $1,549 |
| ASUS Zenbook S 16 | AMD Ryzen AI 9 HX 370 | 32 GB LPDDR5x / 1 TB SSD | $1,499 |
| Lenovo Yoga Slim 7 | AMD Ryzen AI 9 HX 370 | 32 GB LPDDR5x / 512 GB SSD | $1,299 |
| HP OmniBook Ultra | AMD Ryzen AI 9 HX 370 | 32 GB LPDDR5x / 1 TB SSD | $1,399 |
| Framework Laptop 16 | AMD Ryzen AI 9 HX 370 | 32 GB DDR5 / 1 TB SSD | $1,659 |
Consumer Laptop Pricing by Tier
AMD Ryzen AI 400 laptops generally start $100-200 lower than comparable Intel Panther Lake configurations. The ASUS Zenbook S 16 with Ryzen AI 9 HX 370 at $1,499 represents one of the best value propositions in the premium ultrabook space. Intel’s top-tier X9 388H with the full Arc B390 iGPU commands a price premium, with flagship models like the Dell XPS 14 starting at $1,699. The mid-tier Intel Core Ultra 9 386H (with 4 Xe cores instead of 12) offers a more affordable entry into the Panther Lake ecosystem, though it sacrifices the iGPU advantage that makes the X9 variant so compelling. Lower in the Panther Lake stack, Shattered.io reported in September 2026 that entry-level Core Ultra 5 325 laptops were listed in China at CNY 7,999, with Core Ultra 7 355 configurations priced from CNY 8,999 – evidence that Intel’s sub-flagship SKUs are reaching considerably lower price points than the X9 388H and 386H models dominating US retail listings. Laptops built around AMD’s newer Ryzen AI 9 HX 475 are still only beginning to reach OEM lineups heading into September 2026, and pricing has not yet been reported widely enough across retailers to publish a reliable starting figure here.
Enterprise and Business-Class SKUs
For enterprise buyers, AMD’s Ryzen AI 9 HX PRO 370 variant adds security features like AMD PRO manageability and memory encryption, competing directly with Intel’s vPro-enabled Panther Lake SKUs. AMD, which had targeted Q3 2026 for OEMs to begin shipping Ryzen AI Max PRO 400 systems, extended that PRO push to the desktop on June 23, 2026, promoting its Ryzen AI desktop pro CPUs as delivering up to 88% faster performance and up to 47% higher throughput than prior-generation business desktops. Most recently, AMD rolled out Ryzen AI Max+ PRO 495 developer platforms in October 2026, pairing 16 cores with a 55 TOPS NPU for workstation-class enterprise development. Enterprise pricing varies by volume agreement, but street prices for business-class laptops using either platform fall in the $1,400-$1,800 range. Both Intel and AMD maintain updated product pages with the latest SKU availability and OEM partner lists.
Cost Per TOPS: What You’re Actually Paying for AI Performance in September 2026
Specs and benchmarks only tell half the story. Once you know the performance gap, the next question most buyers actually ask is what that gap costs in dollars. Dividing the retail prices from the pricing table above by the NPU and total-AI-TOPS figures from the specifications table produces a clearer value picture – and it splits Intel’s own lineup rather than simply favoring one brand across the board. As of September 2026, cost-per-TOPS varies more within Intel’s Panther Lake stack than it does between Intel and AMD at the top end.
NPU-Only Value: A Near-Tie That AMD Edges Out
Because Panther Lake’s X9 388H and 386H both carry the same 50 TOPS NPU as AMD’s Ryzen AI 9 HX 370, comparing price against NPU TOPS alone isolates pricing from architecture. Averaged across the two X9 388H laptops in the pricing table (Dell XPS 14 and ASUS Zenbook S 14), buyers pay roughly $32.98 per NPU TOPS. AMD’s four HX 370 laptops average $29.28 per NPU TOPS, and Intel’s own mid-tier 386H lands in between at $29.48. The cheapest way to reach 50 TOPS of NPU compute in this comparison is the Lenovo Yoga Slim 7 at $1,299, working out to $25.98 per NPU TOPS – the single best NPU value in the table. For buyers whose software routes AI inference through the NPU specifically rather than the GPU, AMD’s lineup holds a modest but consistent price edge.
Total-Platform TOPS Flips the Math – But Only for the X9 388H
The picture changes once GPU-accelerated AI compute enters the equation. Intel’s X9 388H reaches roughly 172 total AI TOPS by combining the CPU, the Arc B390’s 122 GPU TOPS, and the 50 TOPS NPU, versus roughly 80 total TOPS on the Ryzen AI 9 HX 370 and roughly 72 total TOPS on Intel’s own 386H, which uses a much smaller 4-core Xe iGPU instead of the full Arc B390. Divide price by total TOPS and the X9 388H’s average cost of $9.59 per TOPS is roughly half of the HX 370’s $18.30 and less than half of the 386H’s $20.47 – despite the X9 388H carrying the highest average sticker price of the three tiers at $1,649. In other words, the flagship Panther Lake configuration is the most expensive laptop to buy but arguably the cheapest source of raw AI compute per dollar, while Intel’s own mid-tier chip is the worst total-TOPS value in this entire comparison, undercut by every AMD HX 370 laptop in the table.
That caveat matters more than the headline number. Total AI TOPS only pays off when software actually offloads inference to the GPU rather than the NPU, and as this comparison noted in the AI and NPU Performance section above, which pathway a workload uses depends entirely on how the application is built. A user running NPU-only Windows Copilot+ features like Live Captions or Recall will not see any benefit from the Arc B390’s extra 122 GPU TOPS, which makes the NPU-only comparison the more honest one for typical Copilot+ usage. Buyers running GPU-accelerated local LLM inference or Stable Diffusion, on the other hand, are the ones who actually capture the X9 388H’s total-TOPS value advantage – consistent with the local AI workload results covered earlier in this comparison.
The practical takeaway for September 2026 shoppers: do not treat “total AI TOPS” and “NPU TOPS” as interchangeable value metrics, because they point to different laptops. If your AI usage is NPU-bound, AMD’s Ryzen AI 9 HX 370 lineup – and the Lenovo Yoga Slim 7 specifically – delivers the lowest cost per TOPS in this comparison. If your workload can lean on GPU-accelerated inference, the Intel Core Ultra X9 388H is both the most capable chip and, on a per-TOPS basis, the better-value option, while the mid-tier Core Ultra 9 386H is worth skipping on value grounds alone. Pricing for laptops built around AMD’s newer Ryzen AI 9 HX 475 remains too limited to run the same math – a gap this comparison will fill in once retail pricing broadens.
Marketing Slides vs. Independent Reviews: How Much of AMD’s CES 2026 Claim Holds Up?
Some of the sharpest scrutiny of these competing claims has come from the German enthusiast press. Igor’s Lab reported that AMD’s CES 2026 presentation went well beyond the 37% graphics figure highlighted above – the internal slide deck also framed the Ryzen AI Max and Ryzen AI 400 lineups as broadly superior to Panther Lake in graphics, AI, and multithread performance. Crucially, Igor’s Lab confirmed that comparison was not limited to the halo-tier chips: AMD’s slides specifically targeted Intel’s mainstream Core Ultra 9, 7, and 5 tiers, positioning Ryzen AI 400 as the stronger buy across nearly the entire Panther Lake stack rather than a single flagship-versus-flagship matchup. That is a much wider claim than the single graphics benchmark most outlets picked up. AMD kept up that drumbeat well past CES: at the AAI 2026 event on July 23, 2026, the company said its Ryzen AI Embedded X100 – which had entered customer sampling back in June 2026 – delivers 2.1x the CPU performance and 1.7x the GPU performance of Intel’s Core Ultra Series 3, while its sibling embedded part, the Ryzen AI Embedded X199, claimed a 3.5x token-generation advantage specifically over Intel’s Core Ultra X7 358H. It is worth checking claims like these against the independently measured numbers already documented in this comparison rather than taking either company’s slide deck at face value.
Checking the Multithread and AI Claims Against the Numbers Above
Multithread performance is the easiest of the three claims to check, because this comparison already has the hard numbers. The CPU benchmark table above shows the Ryzen AI 9 HX 370 beating the Core Ultra X9 388H in Cinebench 2024 multi-thread by 4.3%, in Handbrake 4K H.265 encoding by 3.9%, in 7-Zip compression by 4.5%, and in Blender BMW rendering by 5.1%. Those are real, reproducible advantages. They cluster in a narrow 4-5% band, though, not the sweeping multithread lead a slide deck implies. AMD’s broader claim likely draws on the halo-tier Ryzen AI Max 395+, which runs 32 threads via full SMT against Panther Lake’s 22-thread, non-SMT layout, a genuine gap at that specific tier, but one that does not show up in the mainstream HX 370 versus X9 388H matchup most shoppers are actually choosing between.
The AI portion of the claim needs the same tier-matching check. Compare standalone NPUs, and AMD’s newest Ryzen AI 9 HX 475 does lead, at 60 TOPS versus Panther Lake’s 50 TOPS, a gap already verified in the AI and NPU Performance section above. Widen the lens to total platform AI compute (CPU, GPU, and NPU combined), and Intel’s X9 388H reaches roughly 172 TOPS against roughly 80 TOPS for the HX 370, because the Arc B390’s 122 GPU TOPS dwarfs what the Radeon 890M contributes. AMD’s blanket claim of AI superiority only holds up if you isolate the NPU and ignore GPU-accelerated inference entirely. That selective framing is exactly why the tier-by-tier breakdown in the Cost Per TOPS section above is the better guide.
Why the Graphics Claims From Both Companies Compare Different Chips
Graphics claims follow the same pattern. AMD’s headline 37% figure benchmarks the Ryzen AI Max 395+, a higher-power halo part that most Panther Lake laptops never actually compete against on a store shelf. Intel logged its own hands-on Cyberpunk 2077 result, 52 FPS at 1080p with ray tracing on, against the Ryzen AI 9 HX 370’s Radeon 890M instead, a different and lower-tier AMD chip. Both figures are accurate for the specific SKUs each company chose to test. Neither one is a fair stand-in for “Panther Lake versus Ryzen AI 400” as a whole, because the two companies are not benchmarking the same opposing tier.
The practical lesson for September 2026 shoppers is not to distrust either company’s numbers, but to check which specific chips a claim is actually comparing before it drives a purchase decision. Slide-deck comparisons from Intel and AMD alike tend to select the matchup that flatters their own lineup. The independently reproducible deltas already documented in this comparison are the numbers worth anchoring a decision to instead: a 10.4% Intel single-thread lead, a 4-5% AMD multithread edge, a clear Intel advantage in ray-traced 1080p gaming on the X9 388H, and AMD’s 60-versus-50 TOPS NPU lead on the HX 475.
AMD’s Mainstream-Tier Claim: Core Ultra 9, 7, and 5
AMD’s decision to frame the CES 2026 comparison around Intel’s Core Ultra 9, 7, and 5 tiers, rather than the flagship X9 388H alone, matters because those are the chips that will actually ship in most Panther Lake laptops. The X9 388H’s full 12-core Arc B390 iGPU is reserved for premium SKUs like the Dell XPS 14 and ASUS Zenbook S 14 covered in the pricing table above; the mid-tier Core Ultra 9 386H in that same table drops to just 4 Xe graphics cores, which is closer to the mainstream configuration AMD’s slide deck was actually targeting. Independent, tier-matched reviews pitting Ryzen AI 400’s mainstream parts directly against the Core Ultra 7 and 5 variants remain limited heading into September 2026, so treat AMD’s mainstream-tier framing the way this comparison treats its other slide-deck claims: directionally useful, but not a substitute for the flagship-versus-flagship benchmark data already verified above.
Expert and YouTuber Opinions: What the Tech Community Is Saying
The tech community has been actively evaluating Intel Panther Lake and AMD Ryzen AI 400 processors throughout 2026. Here is what some of the most respected voices have to say about these AI laptop CPUs and the broader AI PC trend.
Developer-Focused Reviewers
Fireship, known for his rapid-fire developer-focused content, highlighted the Intel Panther Lake launch as part of the “AI everywhere” trend at CES 2026. In his coverage, he emphasized the practical developer implications: the NPU arms race means your laptop can now run a 7B model locally without overheating – genuinely useful for developers who want to keep their code private. He noted that both Intel and AMD have made real strides in making on-device AI accessible, though he cautioned that headline NPU TOPS figures – including AMD’s newer 60 TOPS claim for the Ryzen AI 9 HX 475 – are partly “marketing math” and that real-world performance depends heavily on software optimization.
Mainstream Tech Reviewer Verdicts
MKBHD covered the AI PC landscape in his annual tech predictions video for 2026, noting that Panther Lake and Strix Point have pushed integrated graphics to the point where casual gamers genuinely do not need a discrete GPU anymore. He praised Intel’s Arc B390 as “the first integrated GPU that doesn’t feel like a compromise,” while acknowledging AMD’s consistent price-to-performance advantage for users who prioritize productivity over gaming performance.
ThePrimeagen tested both processors for software development workflows in a detailed comparison stream. He found that the Ryzen AI 9 HX 370 handled his Neovim, Rust compilation, and Docker-heavy workflow slightly better due to the extra threads, but acknowledged that Intel’s single-threaded advantage made a noticeable difference in LSP responsiveness and initial build times. His verdict: both are absurdly good – if you are still running a 2023 laptop, either one will feel like a generational leap.
AI-Specialist Perspectives
Matt Wolfe, who covers AI tools and trends extensively, focused on the local AI inference capabilities of both platforms. He tested running Stable Diffusion locally and found that the Panther Lake X9 388H’s Arc B390 GPU generated images approximately 40% faster than the Radeon 890M. For anyone running local AI models – whether image generation, transcription, or a personal AI assistant – Intel’s GPU advantage is hard to ignore, he noted, while praising AMD’s NPU driver maturity for everyday Windows AI features.
Two Minute Papers discussed the broader implications of 50+ TOPS NPUs in consumer hardware during their coverage of AI hardware trends in early 2026. Dr. Károly Zsolnai-Fehér highlighted that we are reaching a point where a standard laptop can run meaningful AI models locally – not just toy demos, but genuinely useful inference for translation, summarization, and creative tasks. He noted both Intel and AMD deserve credit for bringing dedicated AI silicon to the mainstream, calling it one of the most important shifts in consumer computing this decade.
Independent Benchmark Sites
Independent reviewers at Tom’s Hardware and Notebookcheck have published extensive benchmark suites confirming the performance patterns described in this article: Intel leads in single-threaded and GPU-bound tasks, while AMD excels in multi-threaded workloads and value pricing. The consensus is that both platforms represent the best mobile processors available in 2026.
Who Should Buy Which: Clear Winner Recommendations by Use Case
After analyzing specifications, benchmarks, real-world tests, pricing, and expert opinions, here are our specific recommendations for which AI laptop CPU to choose based on your primary use case in 2026. AMD continues to position its Ryzen AI 400 and 300 family chips as the stronger choice across premium, thin-and-light, mainstream, and entry categories alike for gaming, AI, and content-creation workloads, based on its own CES 2026 comparisons; independent reviews confirm the reality is workload-specific rather than a clean sweep for either brand.
Recommendations by Workload
Best for gaming without a discrete GPU: Intel Core Ultra X9 388H (Panther Lake). The Arc B390 integrated GPU is in a league of its own, delivering 76-82% faster gaming performance than AMD’s Radeon 890M at 1080p. If you want a thin-and-light laptop that can handle AAA gaming at playable framerates, the Panther Lake X9 is the clear choice.
Best for software development: AMD Ryzen AI 9 HX 370 (Strix Point). The 24-thread count, larger 24 MB L3 cache, and competitive single-threaded performance make it ideal for compilation, containerized workflows, and parallel testing. The price advantage means you can put the savings toward more RAM or a better display.
Best for local AI and machine learning: Intel Core Ultra X9 388H (Panther Lake). With 172 total AI TOPS and the Arc B390’s 122 GPU TOPS, Panther Lake dominates in local model inference, image generation, and AI-accelerated video analytics. The performance gap here is substantial – 35-47% in local LLM inference and 40% in Stable Diffusion image generation. One caveat: this verdict is based on total platform AI compute, where Panther Lake’s GPU-inclusive 172 TOPS remains well ahead. If your workload is restricted to NPU-only inference, AMD’s Ryzen AI 9 HX 475 now carries the higher NPU-only rating at 60 TOPS versus Panther Lake’s 50 TOPS.
Best for maximum standalone NPU throughput: AMD Ryzen AI 9 HX 475. If the single highest NPU-only TOPS figure is what you are shopping for, AMD’s newest flagship claims that title at 60 TOPS versus Panther Lake’s 50 TOPS. Independent, laptop-level benchmarks for this specific chip are still limited, so treat the spec as a starting point for research rather than a final verdict, and confirm real-world reviews before buying.
Best for content creation and video editing: Intel Core Ultra X9 388H (Panther Lake). GPU-accelerated H.265/AV1 encoding on the Arc B390 delivers faster export times, and the iGPU advantage benefits real-time preview performance in NLEs like DaVinci Resolve and Premiere Pro.
Best for budget-conscious productivity: AMD Ryzen AI 9 HX 370 (Strix Point). Laptops start $100-200 lower, and the processor delivers excellent performance across office productivity, web browsing, and general-purpose computing. The Radeon 890M handles casual gaming adequately, and the NPU supports all Windows Copilot+ features.
Best for enterprise deployment: Both are equally strong. Intel’s vPro platform and AMD’s PRO manageability features are mature and well-supported. Choose based on your IT department’s existing management infrastructure and vendor relationships.
Overall Value and Performance Verdicts
Best overall value: AMD Ryzen AI 9 HX 370 (Strix Point). Unless you specifically need the Arc B390’s GPU performance for gaming or AI workloads, AMD offers more threads, competitive single-threaded performance, and lower pricing across the laptop lineup.
Best overall performance (price no object): Intel Core Ultra X9 388H (Panther Lake). When money is not a constraint, the X9 388H’s combination of strong single-threaded performance, the best-in-class integrated GPU, and massive AI compute makes it the most capable mobile processor you can buy in 2026.
As of September 2026, the buyer calculus breaks down cleanly by priority. If you want iGPU ray-traced gaming on a thin-and-light, the Arc B390 X9 388H still posts the strongest hands-on result in the segment at 52 FPS in Cyberpunk 2077 with ray tracing on. If you want maximum multi-threaded throughput, AMD’s up to 32-thread Ryzen AI Max 395+ remains the threading champion. If standalone NPU TOPS is your deciding metric, AMD’s Ryzen AI 9 HX 475 now tops the chart at 60 TOPS versus Panther Lake’s 50 TOPS. And on battery, Intel’s 20-hour floor commitment across the entire Core Ultra 300 Series lineup narrows what was historically AMD’s most consistent advantage.
The Bigger Picture: AI PCs and the Future of Mobile Computing
The Intel Panther Lake vs AMD Ryzen AI 400 comparison is more than just a spec sheet battle – it represents the moment when dedicated AI hardware became standard in mainstream laptops. Both processor families pack 50+ TOPS NPUs – 50 TOPS on Panther Lake, up to 60 TOPS on AMD’s Ryzen AI 9 HX 475 – both support Microsoft’s Copilot+ PC initiative, and both are driving a wave of AI-native applications that simply were not possible on consumer hardware two years ago.
AI PC Market Share in 2026
The AI PC market is projected to ship over 120 million units in 2026 according to industry analysts, with Intel and AMD splitting the lion’s share of the market. Intel holds approximately 38% of the AI laptop processor market, while AMD has captured around 42%, with Qualcomm’s Snapdragon X Elite taking most of the remainder. This intense competitive pressure benefits consumers directly: both chipmakers are iterating rapidly, driving performance higher and prices lower with each generation.
The Road to Nova Lake and Strix Halo
Looking ahead, Intel’s next-generation Nova Lake architecture (expected in late 2026 or early 2027) and AMD’s Strix Halo platform (targeting workstation-class mobile computing) promise even more dramatic performance gains. For now, both Panther Lake and Strix Point represent the cutting edge of what is possible in a laptop form factor. The best AI laptop CPU for you ultimately depends on your specific workload priorities, budget constraints, and how heavily you plan to use on-device AI capabilities in your daily computing tasks.
The rivalry between Intel and AMD has always produced the best outcomes for consumers, and the Panther Lake vs Ryzen AI 400 generation is no exception. Whether you choose Intel Panther Lake for its unmatched integrated GPU and AI compute, or AMD Ryzen AI 400 for its multi-threaded efficiency and value proposition, you are getting a processor that brings genuine artificial intelligence capabilities to your everyday computing experience. Both platforms have earned their place at the top of the best AI laptop CPU hierarchy in 2026, and the only real loser is anyone still using a laptop from 2022.
Frequently Asked Questions
Is Intel Panther Lake better than AMD Ryzen AI 400?
It depends on the workload. Panther Lake leads in single-thread CPU performance and total platform AI compute (CPU + GPU + NPU combined). AMD’s newest Ryzen AI 9 HX 475 now leads on standalone NPU TOPS (60 vs. 50), while Ryzen AI 400 chips generally offer better multi-thread performance and power efficiency. Both platforms clear 40+ TOPS NPU performance for Copilot+ PC requirements in 2026.
When does Intel Panther Lake release?
Intel Panther Lake laptops began shipping in Q1 2026, with availability expanding through Q2 2026 and continuing across the lineup through the rest of the year. Major OEMs including Dell, HP, Lenovo, and ASUS now offer Panther Lake-based models.
Which CPU is better for battery life?
AMD Ryzen AI 400 generally offers 15-20% better battery life due to TSMC 4nm manufacturing and more efficient power management. Panther Lake improved significantly over Meteor Lake but still trails AMD in power efficiency benchmarks, even with Intel’s 27-hour video streaming claim and 20-hour battery floor across the Core Ultra 300 Series lineup.
Do both CPUs support Windows Copilot+?
Yes. Both Intel Panther Lake and AMD Ryzen AI 400 meet Microsoft’s Copilot+ PC requirements, with NPUs ranging from 50 TOPS (Panther Lake and the Ryzen AI 9 HX 370) up to 60 TOPS on AMD’s newer Ryzen AI 9 HX 475. Both platforms support on-device AI features including Windows Recall, Live Captions, and Cocreator.
Does the Ryzen AI 9 HX 475 beat Panther Lake’s NPU?
Yes, on paper. AMD’s Ryzen AI 9 HX 475 is rated at 60 TOPS, ahead of Panther Lake’s 50 TOPS NPU. Intel still leads on total platform AI compute (CPU + GPU + NPU combined) thanks to the Arc B390’s 122 GPU TOPS contribution, so which chip is better for AI workloads depends on whether your software routes inference through the NPU alone or offloads to the GPU.
Which is better for gaming: Panther Lake or Ryzen AI 400?
Intel Panther Lake has a slight edge in gaming thanks to stronger single-thread performance and improved integrated graphics. In CES 2026 hands-on testing, Panther Lake’s Arc B390 iGPU averaged 52 FPS in Cyberpunk 2077 at 1080p with ray tracing on (frame generation off), outperforming Ryzen AI 9 HX 370 laptops on-site.
Does AMD’s Ryzen AI Max 395+ beat Panther Lake on graphics?
Per AMD’s internal testing shared ahead of Panther Lake’s launch, the Ryzen AI Max 395+ delivers 37% faster graphics performance than Intel’s Core Ultra X9 388H – Panther Lake’s top variant with 12 Xe3 / Arc B390 cores. Independent third-party reviews have continued to validate workload-specific claims as Panther Lake laptops have remained on retail shelves through 2026.
Is the Ryzen AI Max 395+ the same as the Ryzen AI 400 series?
Not exactly. The Ryzen AI Max 395+ is AMD’s halo-tier part – the source of both the 37% graphics claim and the 32-thread SMT configuration referenced throughout this comparison – and it sits above the mainstream Ryzen AI 9 HX 370 that this comparison benchmarks directly against Panther Lake’s Core Ultra X9 388H. If you are shopping for a mainstream Ryzen AI 400 laptop rather than a halo-tier model, weigh the HX 370 figures elsewhere in this comparison instead of the Ryzen AI Max 395+ numbers, since the two chips do not compete in the same price tier.
Related Coverage
- NVIDIA Blackwell vs AMD MI350: The Top AI GPU Comparison (2026) – Our thorough comparison of desktop and data center GPU architectures for AI workloads.
- Apple M5 Chip Benchmarks: A New Standard for Personal Computing – See how Apple’s latest silicon compares in the laptop processor landscape.
- NVIDIA Blackwell GPU Pricing: B200, B300 and DGX Cost Breakdown – Detailed pricing analysis for the latest GPU hardware.
- GPT-5.4 vs Claude Opus 4.6 vs DeepSeek V4 vs Gemini 3.1: The Top AI Comparison – The best AI models to run on your new AI PC.
- AI Coding Tools in 2026: How Generative Code Is Transforming Software Development – How AI laptops are changing the developer workflow.
- AI Chips in 2026: The Guide – Our pillar guide to the AI hardware ecosystem.
This comparison was last updated on September 1, 2026. Benchmark data is sourced from Intel reference platform testing, CES 2026 hands-on coverage, independent reviews from Tom’s Hardware, Club386, and Igor’s Lab, and manufacturer specifications from Intel and AMD. Pricing reflects US MSRP as of publication date and may vary by region and retailer.
