NVIDIA announced the GeForce RTX 50 series back on January 6th at CES 2025 and since then we’ve desperately waited for benchmark data. It looks like we’re getting our first batch of data now with some decent (but not great) numbers showing the performance enhancements.
These benchmarks suggest performance improvements ranging from 15% to 33% compared to the RTX 40 series. It’s important to note that these figures are based on Nvidia’s internal testing. Real-world performance can vary based on the game, settings, and resolution used.
RTX 50 Series Performance: A Closer Look
RTX 50 Series Benchmark Data
It’s crucial to understand that at this pre-release stage, concrete, independently verified benchmark data for the RTX 50 series is limited. Nvidia has released some internal benchmark figures, but these should be taken with a degree of caution until third-party testing is available. The information below is based on Nvidia’s provided data and leaks, and should be considered preliminary.
Preliminary Benchmark Estimates
The table below summarizes the available information. Note that these are estimates and subject to change:
| Card | Estimated Performance Increase (vs. Previous Gen, Rasterization) | Notes |
|---|---|---|
| RTX 5090 | Up to 33% | This is the highest reported gain, likely in ideal conditions. |
| RTX 5080 | Around 15% | This is a significantly smaller jump than the 5090 and has caused some discussion. |
| RTX 5070 Ti | Around 20% | Appears to offer a good balance of price and performance. |
| RTX 5070 | Around 20% | May perform similarly to the RTX 4070 Super. |
Important Considerations:
- These figures are based on Nvidia’s internal testing and leaks. Independent benchmarks are needed for confirmation.
- The performance increase will vary depending on the game, resolution, and settings used.
- The introduction of DLSS 4 and Multi Frame Generation makes direct FPS comparisons more complex.
- It is important to remember that these are preliminary numbers and can change.
Understanding the Benchmarks
Nvidia has released initial benchmark data for its upcoming RTX 50 series graphics cards, which are based on the Blackwell architecture. By utilizing technologies such as DLSS 4.0, some models may experience a potential performance boost of up to 33% in certain games. While early benchmarks indicate performance improvements, the overall value of the RTX 50 series is still being evaluated.
This article will compare the new series with previous generations and AMD products, taking into account factors like benchmarking methods, power consumption, and pricing to assist in purchasing decisions. It’s important to note that the 33% gain is likely a best-case scenario, and the performance enhancement varies among specific cards in the lineup. For instance, the performance jump from the 4080 to the 5080 appears smaller than the jump from the 4090 to the 5090. This trend has been observed in previous generations as well, with top-tier cards typically showing the most significant gains.
Multi Frame Generation (MFG) and Benchmarking Changes
Nvidia’s DLSS 4 includes Multi Frame Generation (MFG). This technology creates entirely new frames using AI, increasing the perceived frame rate. This new technology requires changes to how we measure performance. Standard FPS (frames per second) measurements become less accurate, as some frames are generated rather than fully rendered. Nvidia recommends using MsBetweenDisplayChange, a timing metric, for more accurate measurements with MFG. This metric captures dropped frames and variations in frame times more effectively.
This also means that comparing raw FPS numbers between cards with and without MFG is not a direct comparison. A game running at 100 FPS with MFG won’t feel the same as a game running at 100 FPS natively rendered. MFG introduces some latency. This latency can be higher with some MFG settings.
Comparing RTX 50 Series to RTX 40 Series and AMD Alternatives
Based on the available information, here’s a comparison of the RTX 50 series to the previous generation and AMD’s competing cards:
| Card | Estimated Performance Increase (vs. Previous Gen) | Key Features/Considerations |
|---|---|---|
| RTX 5090 | Up to 33% | Top-tier performance, largest generational leap. |
| RTX 5080 | Around 15% | Smaller performance increase compared to the 5090; may not significantly outperform the 4090 in all cases. |
| RTX 5070 Ti | Around 20% | Good balance of price and performance, more VRAM than 4070 Ti. |
| RTX 5070 | Around 20% | May offer similar performance to the 4070 Super. |
| AMD Radeon RX 7900 XTX | N/A | A strong competitor to the RTX 4080 and potentially the RTX 5080 in raw performance, especially in some rasterization workloads. |
Several sources suggest that the performance increase for the 5080 is smaller than many hoped. This has led to discussion about whether the 5080 offers good value compared to the 4090 or AMD’s 7900 XTX. The 5070 Ti is generating some interest, as it appears to offer a decent performance jump with a good amount of VRAM at a reasonable price, especially compared to the 4070 Ti it replaces. The 5070 also seems to be a good card, but whether it is worth it over a discounted 4070 Super is up for debate.
Power Consumption and Efficiency
The RTX 50 series, particularly the high-end cards, are expected to have higher power consumption than their predecessors. The 5090, for example, has a higher power limit than the 4090. This increase in power consumption is a trade-off for the performance gains. However, some early analysis suggests that the performance per watt may remain similar or improve slightly.
The Role of DLSS and Ray Tracing
DLSS (Deep Learning Super Sampling) and ray tracing continue to be important features for Nvidia GPUs. DLSS uses AI to upscale lower-resolution images to higher resolutions, improving performance without significant loss of visual quality. Ray tracing creates more realistic lighting and reflections in games. The RTX 50 series is expected to improve ray tracing performance and introduce new DLSS technologies, like DLSS 4 with MFG. These technologies will be a key differentiator from AMD’s offerings.
Considering an Upgrade?
If you own an RTX 4090, an upgrade to the 50 series may not be necessary unless you need the absolute top performance or want to take advantage of the latest DLSS features. If you have a lower-tier 40 series card or an older GPU, the 50 series might be a worthwhile upgrade, depending on your budget and needs. The 5070 Ti seems to be a good choice for those looking for a balance of price and performance. It is important to wait for independent reviews once the cards are released to get a clearer picture of their real-world performance.
Beyond Gaming: AI and Professional Workloads
While gaming performance is a primary focus, Nvidia GPUs are also used for AI development, content creation, and other professional workloads. The RTX 50 series, with its improved architecture and increased memory bandwidth, is expected to offer significant improvements in these areas as well. The increased memory capacity on some cards is especially beneficial for AI tasks that require large datasets.
Understanding GPU Architectures and Node Processes
The performance of a graphics card depends on several factors, including its architecture and the manufacturing process used to create its chips. The RTX 50 series is based on Nvidia’s Blackwell architecture, which is designed to improve performance and efficiency. The manufacturing process, often referred to as the “node,” refers to the size of the transistors on the chip. Smaller nodes generally allow for more transistors to be packed onto the same chip area, leading to increased performance and reduced power consumption.
However, the naming conventions for these nodes (like 4nm or 3nm) are primarily marketing terms and don’t reflect the actual physical measurements of the transistors. The transition from one node to another doesn’t always result in a proportional increase in performance or transistor density. Other architectural improvements also play a significant role. This is why the performance gains from the RTX 40 to the RTX 50 series are not as large as some previous generational jumps.
Short Summary:
- NVIDIA states that its new RTX 50 series can perform up to twice as fast as the previous generation.
- DLSS 4.0 provides major enhancements in frame generation, increasing overall gaming performance.
- Technical specifications reveal significant upgrades in hardware and architecture across the lineup.
The flagship model, RTX 5090, along with RTX 5080, RTX 5070, and RTX 5070 Ti, is scheduled for release between January and February this year. According to NVIDIA’s CEO, Jensen Huang, the RTX 5090 can deliver up to double the performance compared to the RTX 4090. This impressive performance gain is largely attributed to the new DLSS 4 technology.
The unveiling of this model highlights a significant increase in hardware capabilities. The RTX 5090 is built on NVIDIA’s advanced Blackwell architecture, which substantially enhances its performance metrics. A notable improvement is the increase in the number of CUDA cores, which has risen from 128 in the RTX 4090 to 170 in the RTX 5090. This 33% increase in core count translates to better overall efficiency and processing power.
Additionally, the memory bandwidth has also received a significant upgrade. The new GDDR7 memory operates on a 512-bit bus clocked at 28 Gbps, offering an impressive 1,792 GB/s of bandwidth—almost an 80% enhancement compared to the RTX 4090.
However, skepticism has arisen regarding these ambitious claims. As Slashdot reader jjslash noted, “NVIDIA has historically leveraged marketing tactics to amplify performance claims.” While the RTX 50 series GPUs showcase considerable improvements through DLSS 4, which enables multi-frame generation not seen in earlier models, these advancements do not always guarantee improved game latency or overall playability.
Initial analyses suggest that the RTX 5070 may actually match the RTX 4090 in frame rates under optimal conditions, particularly when utilizing DLSS 4. Nonetheless, it is essential to evaluate these performance metrics separately from the raw, non-DLSS performance interactions.
“The claims surrounding the RTX 5070 delivering performance akin to the RTX 4090 may be overstated,” noted an industry analyst. “DLSS 4 is a game-changer, but without it, you’re unlikely to see those numbers hold true.”
NVIDIA furnished benchmarks where the RTX 5070 appears to offer a performance uplift of approximately 20% to 30% over the RTX 4070 at 1440p resolutions, alongside a competitive price tag of around $550, which is slightly less than the RTX 4070 Super. These adjustments could position the RTX 5070 as an appealing alternative for gamers seeking performance without emptying their wallets.
A critical aspect overshadowing the performance leaps offered by DLSS 4 is the technical challenge posed by latency. Despite the impressive raw FPS numbers that DLSS technology may exhibit, continued usage indicates occasional latency that could hinder certain gameplay experiences. “You might achieve a jump from 30 frames per second to 120 frames, but that doesn’t change the intrinsic responsiveness you get from a low base frame rate,” warned a tech reviewer familiar with NVIDIA’s technology.
Technical Enhancements: What’s New?
Beyond performance claims, the engineering behind NVIDIA’s RTX 50 series reveals some fascinating improvements. Each GPU in the series has undergone significant hardware enhancements:
- CUDA Cores: The RTX 5090 features 170 CUDA cores, a jump from 128 formatted in its predecessor.
- Memory Bandwidth: With GDDR7 technology, the RTX 5090 delivers an impressive bandwidth of 1,792 GB/s, nearly 80% stronger than the RTX 4090.
- VRAM Capacity: For the first time, consumer GPUs in this series are equipped with an expansive 32GB of VRAM.
This series isn’t just about raw performance; it’s also about environmental sustainability. The RTX 50 graphics cards have been designed keeping efficiency in mind, promising reduced energy consumption while significantly boosting performance. Analysts hailed NVIDIA’s strategy, suggesting that it may set a precedent for future GPU designs focused on eco-friendliness.
However, enthusiasts are advised to temper their excitement, as early benchmarks presented by NVIDIA during the unveiling may not reflect real-world gaming performance. A tech expert referred to the numbers as “preliminary,” recommending gamers and professionals alike to await independent benchmark tests to validate the marketing. “You have to approach these numbers with a grain of salt,” he cautioned.
DLSS 4: The Game-Changer?
The introduction of DLSS 4 technology heralds a new era of graphical fidelity and performance in gaming, expanding on the capabilities of DLSS 3. Not only does DLSS 4 provide superior frame rates, but it does so by generating intermediate frames in fast-paced segments of games. Analysts have noted that this could effectively mask frame generation issues while enhancing visual fluidity in compatible titles.
“What we see with DLSS 4 is revolutionary in frame generation, but it remains crucial to evaluate how this impacts player experience with respect to input latency,” explained a performance consultant.
Significantly, DLSS 4 assists in maintaining high frame rates in demanding titles, allowing RTX 50 GPUs to offer smoother gameplay even at higher resolutions. Dual-frame generation considerably reduces strain during high-action sequences, helping players keep pace during crucial moments. However, concerns over potential latency increase persist, suggesting that players might experience momentary gameplay lags that could affect high-speed interactions.
The result? While reports indicate that players can expect remarkable generational improvements, they should also remain conscious of the trade-offs involved. For example, across Nvidia’s tests conducted for the RTX 5090 and RTX 5070, games like “A Plague Tale: Requiem” noted a performance increase of around 43% compared to the RTX 4090 under conditions without DLSS benefits. Comparatively, the RTX 4090 managed to outperform the RTX 3090 by 72% in the same game under similar operational settings, showcasing a disparity between DLSS-backed processing versus intrinsic GPU performance uplift.
Competitive Landscape and Consumer Choice
As the gaming community evaluates these advancements, discussions pivot to NVIDIA’s standing against AMD. Recent reports highlight that the introduction of DLSS 4 is NVIDIA’s answer to AMD’s FSR 4.0, which emphasizes comparative processing capabilities. This competition could propel GPU innovations further and provide gamers with more choices that align with their preferences.
A critical note among industry experts is that the performance advantages of NVIDIA systems must also factor into gamer investment. While the price point ticks upward in the RTX 50 series, the return in gaming capability is redefining industry standards. Some analysts argue that NVIDIA must deliver consistent performance results across the entire range.
“The stakes are high; users expect more than mere percentages in performance upgrades,” said a senior market analyst. “They demand practical applications and enhancements to gaming experiences overall.”
As gaming technology advances, consumers should consider performance metrics and pricing when upgrading hardware. The significance of DLSS and its influence on gameplay is crucial for gamers. NVIDIA’s RTX 50 series, featuring DLSS 4, is generating interest, but thorough reviews and benchmarks are necessary to assess its performance compared to AMD GPUs. The RTX 50 series promises enhanced graphic fidelity and energy efficiency, potentially reshaping consumer expectations. Gamers are advised to remain cautious about claims of performance improvements until verified by third-party sources.
