The strongest available comparison is encouraging precisely because the two HP machines are broadly similar in several fundamentals. Yet it also shows why a single battery-runtime number should not be transformed into a sweeping verdict about an entire processor platform.
The HP result: nearly identical web-browsing endurance
The newer EliteBook X G2i tested with Intel’s Core Ultra 7 356H lasted 17 hours and 59 minutes in a Wi-Fi web-surfing test run at 150 cd/m². The prior-generation EliteBook X G1i with a Lunar Lake Core Ultra 7 268V managed 18 hours and 30 minutes under the corresponding test.
That is a 31-minute advantage for the older system, or roughly 2.8%. In ordinary use, that is close enough that it would be difficult to call the result a decisive practical win for either machine. A person working through web applications, email, documents, and browser-based collaboration tools could reasonably view both results as all-day battery life, subject to their own display brightness, wireless use, background tasks, and workplace software.
The comparison is useful because the reviewed systems shared several important characteristics:
- Both had a 68 Wh battery.
- Both used 1,920 × 1,200, 60 Hz IPS-class displays.
- Both were configured with 32 GB of memory.
- Both were premium HP EliteBook X systems in a comparable portable-business category.
That makes the near tie much more meaningful than comparing unrelated laptops with radically different batteries and screen sizes. It supports a limited conclusion: in this particular HP pairing and in this particular browsing workload, Panther Lake did not require a large battery-life sacrifice to provide higher CPU performance.
It does not establish that every Core Ultra Series 3 laptop will equal a Lunar Lake system in battery life. Nor does it isolate the processor as the reason for the observed result.
Idle power complicates the headline
The same tests show a less flattering result for Panther Lake at idle. At 150 cd/m², the G2i was measured at 3.72 W, while the Lunar Lake G1i measured 3.20 W. The difference is 0.52 W, with the older system using about 16.3% less power in that idle condition.
That figure should not be dismissed as “slight” merely because the absolute wattage difference looks small. At low power levels, fractions of a watt can matter, particularly during long periods of reading, note-taking, messaging, or waiting in a browser. It is also a reminder that similar measured battery life does not necessarily mean identical platform behavior.
The two findings are not inherently contradictory. Idle power is one narrowly defined state; browsing is an active, changing workload that includes display behavior, wireless traffic, memory activity, storage activity, operating-system scheduling, and intermittent CPU work. A laptop can draw more power in one measured idle scenario and still finish a particular web test with a very similar runtime.
For buyers, the lesson is practical: battery-life claims should be tied to a workload. “Up to” figures and broad efficiency labels cannot tell a prospective owner whether their own day looks more like idle time, browser activity, local productivity work, video playback, or sustained CPU tasks.
Panther Lake’s performance lead is real, but its size depends on the test
The Core Ultra 7 356H is a launched Core Ultra Series 3 mobile processor formerly known as Panther Lake. Intel lists it with 16 total cores: four performance cores, eight efficient cores, and four low-power efficient cores. Its listed processor base power is 25 W, while maximum turbo power is 80 W.
Those specifications describe the processor’s available power envelope, not the exact power a particular laptop will consume in daily use. Manufacturer firmware, cooling, power profiles, memory, the screen, and the Windows workload all help determine how a system behaves.
In the HP comparison, the newer G2i held a clear performance advantage. But describing it simply as “32% faster” would be too vague.
Notebookcheck’s composite CPU Performance Rating was 82 for the Panther Lake G2i and 62.2 for the Lunar Lake G1i. That is a 31.8% increase. It is fair to describe Panther Lake as around 32% ahead only when referring specifically to that composite rating.
Other results show why the qualifier is essential:
- Cinebench R23 multi-core measured 17,980 on the G2i and 10,913 on the G1i, a difference of about 64.8%.
- PCMark 10 system score measured 8,393 on the G2i and 7,792 on the G1i, a difference of about 7.7%.
These are not competing truths. They measure different things. A multi-core rendering-style benchmark can emphasize CPU throughput and reward a design with considerably greater sustained parallel performance. A broader system benchmark is influenced by a wider set of factors and may better resemble some everyday desktop tasks, but it is not a universal measure of how fast every Windows application will feel.
The important conclusion is that the G2i’s performance improvement appears especially meaningful in CPU-heavy, multi-threaded work. The benefit may be less dramatic in lighter general-purpose tasks. Users working with demanding local workloads should therefore look beyond a single overall percentage and seek tests relevant to their software.
Why this is not a processor-only experiment
The two EliteBooks are similar enough to make the comparison valuable, but not similar enough to prove a CPU-only efficiency result.
The G2i configuration used LPDDR5X-9600 memory and a 1 TB SSD. The G1i used LPDDR5X-8533 memory and a 512 GB SSD. The newer model also had a SureView IPS display and a different, thicker magnesium chassis, while the older system used an IPS touchscreen. Each difference can affect power use, thermal behavior, performance, or both.
That does not invalidate the battery result. It simply places a boundary around what the evidence can show. It demonstrates the behavior of two complete retail laptop configurations, not the intrinsic efficiency of one CPU measured in isolation from every other component.
This caveat is particularly relevant to premium Windows laptops, where system makers make different choices about panel technologies, battery capacities, cooling limits, firmware tuning, and default power modes. A chip that performs well in a lightly cooled business notebook may behave quite differently in a thinner design, a model with a high-refresh display, or a laptop with a larger OLED panel.
Other early Panther Lake reporting reaches a similarly cautious position. Independent reviewers have described strong efficiency and considerably improved multi-core performance, but they have also noted that direct battery comparisons are often not one-to-one. One tested Panther Lake system used a very different dual-screen, 99 Wh design from the Lunar Lake machines available for comparison. Another review reported a 52% multi-core improvement for a higher-tier Panther Lake chip against a Lunar Lake system, while explicitly warning that the laptops were not exact counterparts.
A separate Core Ultra 7 356H business-laptop review also reported low power use in some scenarios and described the platform as efficient. That is useful supporting context, but it is not a controlled Lunar Lake-versus-Panther Lake test. Different workload methods, displays, batteries, and power settings mean the individual results cannot be merged into a single universal battery-life figure.
What this means for Intel’s efficiency message
Intel’s public positioning for Panther Lake has centered on the prospect of Lunar Lake-like efficiency alongside greater performance. The HP result is favorable evidence for that goal, because it combines almost equal browser endurance with stronger CPU benchmark results in comparable business systems.
Still, the broadest version of the claim remains unproven by this evidence alone.
The original efficiency proposition was not equivalent to a guarantee that every retail Core Ultra Series 3 laptop would deliver the same battery runtime as every Lunar Lake laptop. Battery runtime is a system outcome. It depends on more than the CPU, and even well-matched laptop comparisons cannot eliminate all of those variables.
There is also a difference between sustained processor efficiency and a web-browsing result. Both are relevant, but they answer different questions. A buyer who compiles code, runs local analysis, or performs CPU-intensive creation work may care most about performance at sustained power. Someone whose day is dominated by browser tabs, remote meetings, and office applications may care most about screen draw, wireless behavior, background activity, and idle power.
Intel can fairly point to the G2i result as evidence that the new generation can preserve very long battery life while advancing performance. It cannot, on this basis alone, claim a settled across-the-market efficiency victory.
Buying advice for Windows laptop shoppers
A Core Ultra Series 3 label is a promising starting point, not a complete purchasing decision. The available evidence suggests that a Panther Lake business laptop can be an appealing upgrade for someone who needs more multi-core headroom without obviously giving up browser-oriented endurance. But the individual laptop remains the product being bought.
Before choosing between a Panther Lake and Lunar Lake design, prospective buyers should compare reviews of the exact configuration they plan to purchase and focus on several details:
- Battery capacity and screen type: A 68 Wh battery with a 60 Hz IPS panel will not predict endurance for a smaller battery, a brighter panel, or a different display technology.
- Test category: Look separately at web browsing, video playback, idle draw, and sustained productivity or CPU testing. One result cannot represent all of them.
- Performance mode and cooling: Processor specifications include wide power ranges, and laptop makers can tune their designs differently.
- Memory and storage configuration: These can vary even within the same model line and can affect both responsiveness and power behavior.
- Your software: A large multi-core lead is more valuable for workloads that can use it than for light tasks already completed quickly on the older system.
For organizations refreshing Windows fleets, the evidence also argues for pilot testing rather than buying on processor branding alone. Testing the actual selected image, endpoint-security stack, browser policy, conferencing tools, display settings, and docking patterns will provide more actionable battery data than a single synthetic claim. The G2i and G1i comparison shows why: very similar-looking laptops can produce a near battery-life tie while still differing materially in idle draw and workload-specific speed.
A promising result, not a final verdict
Panther Lake’s showing in the two HP EliteBooks is meaningful. With nearly equal 68 Wh batteries, matching resolution and refresh rate, and the same memory capacity, the newer Core Ultra 7 356H system came within 31 minutes of Lunar Lake’s web-surfing runtime while producing much higher CPU-oriented benchmark results.
That is credible evidence that Intel’s newer platform can pair strong multi-core gains with all-day efficiency in a real business laptop. It is not sufficient evidence to declare that Panther Lake has categorically matched Lunar Lake’s efficiency across every laptop design and workload.
The most accurate reading is also the most useful one: Panther Lake looks capable of delivering a substantially better performance-per-battery-life balance in the right Windows laptop. Whether a specific model does so depends on the complete system, and buyers should demand testing that reflects the work they actually do.