NVIDIA launches RTX 5060 and RTX 5060 Ti with up to 16 GB and focus on AI
Table of Contents
After postponing the launch without explanation, the NVIDIA, finally announced the details of the new graphics cards RTX 5060 e RTX 5060Ti, the favorite mid-range option for gamers who don't want to spend a lot but still want access to new features. Find out all the details right now.
What's New in NVIDIA's 50 Series GPUs
Blackwell Architecture
The architecture Blackwell, successor of Ada Lovelace, is the basis for the series video cards 50 da NVIDIA. Developed with cutting-edge technologies, with the manufacturing process TSMC 3sqm, it delivers significant advances in performance and energy efficiency, with greater transistor density and better thermal management.
Improvements include more advanced CUDA cores optimized for parallel computations and superior graphics performance. In addition, the architecture offers improved support for technologies such as Ray Tracing state-of-the-art and greater integration of artificial intelligence, benefiting graphic and computational applications.
With a focus on energy efficiency, Blackwell also delivers higher power per watt, ideal for gamers, content creators and professionals. During the presentation of the NVIDIA na CES 2025, Jensen Huang, president of the company, stated that the most advanced model in the laptop line can achieve the performance of RTX 4090, the most advanced of the last generation, with half the energy consumption.
O Advanced Power Gating turns off inactive parts of the GPU to save power, while the Low Latency Sleep allows the GPU to enter sleep mode frequently, reducing consumption even in use.
DLSS-4
A DLSS-4 (Deep Learning Super Sampling 4) is the next generation of AI-assisted rendering technology from NVIDIA, introduced with GPUs GeForce RTX 50. Its main advancement is Multi-Frame Generation, which allows you to create up to three additional frames for each traditionally rendered frame, significantly increasing performance.
This functionality is exclusive to series GPUs. RTX 50, but other AI model improvements like Super Resolution and Ray Reconstruction are available for all RTX GPUs.
Using an advanced vision transformer architecture, DLSS 4 improves visual quality, with greater temporal stability, less ghosting, and sharper images during motion. NVIDIA claims the technology can increase frame rates by up to eight times compared to traditional methods, enabling 4K gameplay at 240fps with ray tracing complete.
Launched in January 2025 during the passage of NVIDIA na CES 2025, The technology DLSS-4 already supports over 100 games in about three months since it was launched. The company is also working to ensure that new AAA titles also support the technology on the day of their release. The great benefit of the technology DLSS is that the notebook does not need to make an effort to process the images, which allows for lower energy consumption.
Neural rendering
Considered an evolution of DLSS, neural rendering will be used to also prevent the video card from having to work too hard and, naturally, energy consumption will be lower. NVIDIA cited in their press release that they integrated neural networks within programmable shaders to create neural shaders.
RTX Neural Shaders will power the next decade of graphics innovation. They can be used to compress textures by up to 7X, saving massive amounts of graphics memory. They can also be used to create cinematic-quality textures and even more advanced lighting effects in games.
Faces and landscapes will be enhanced to provide higher resolution for users, using only a simpler rasterized face as a base, automatically and without any human intervention. The new architecture's artificial intelligence was essential to make this happen.
GDRR7 memory
With data transfer speeds of up to 36 Gbps per pin, GDDR7 provides much higher bandwidth, essential for 4K gaming, 3D rendering, and intensive applications such as artificial intelligence and machine learning.
One of the most important innovations is the adoption of signage PAM3 (Pulse-Amplitude Modulation 3 levels), which allows for greater efficiency in data transmission compared to previous technologies. This contributes to lower power consumption per bit transferred and helps reduce heat generation, even when operating at high frequencies. In addition, GDDR7 is designed to offer greater density, enabling a larger amount of memory in compact devices.
The memories of the series RTX 50 da NVIDIA are manufactured by Samsung and Micron. The Korean company introduced GDRR7 memory in late 2024 and was expected to actually be helping NVIDIA in the development of its new line of video cards.
Energy Efficiency
Due to the new architecture and also its new memory, the new video cards — both the desktop and notebook versions — will consume less energy to deliver higher performance. And another new feature is also coming to change the field of energy efficiency: the connector fields.
RTX 50 models use the 16-pin connector 12V2×6, a revision of the 12VHPWR connector, designed to improve safety and efficiency in power delivery. This update aims to minimize heat issues and connection failures, contributing to a more stable and efficient operation of GPUs.
High frame rate
The NVIDIA GeForce RTX 5060 Ti, equipped with the Blackwell architecture and 16GB of GDDR7 memory, delivers solid performance for 1080p and 1440p gaming. While it doesn't reach the frame rates of the RTX 5080, 5060 Ti It is an efficient option for gamers looking for graphic quality without compromising their budget. With a 128-bit bus and 200W TGP, it provides a fluid gaming experience in modern titles, especially when combined with technologies such as DLSS and Reflex.
In terms of performance, the RTX 5060 presents notable improvements over its predecessor. Benchmarks indicate that it outperforms the RTX 4060 by up to 32% in tests such as 3DMark Time Spy. This improvement is attributed to the increase in the number of CUDA cores and the adoption of GDDR7 memory, which provides greater bandwidth.
During the presentation event of the new video cards, it was highlighted that the most advanced model (GeForce RTX 5060 TI) can reach up to 234 frames per second when running Hogwarts Legacy in 2K resolution. See the full list:
- Marvel Rivals: up to 330 fps;
- Hogwarts legacy: up to 234 fps;
- Avowed: up to 220 fps;
- Stalkers 2: up to 208 fps;
- Cyberpunk 2077: up to 148 fps;
- Black MythWukong: up to 130 fps;
- Half-Life 2: RTX: up to 130 fps;
- Alan Wake II: up to 114 fps;
- Star Wars Outlaw: up to 105 fps.
GeForce RTX 5060
Simpler version of the intermediate category and of course, successor of the RTX 4060, RTX 5060 presents notable advances over its predecessor. Benchmarks indicate that it outperforms the RTX 4060 by up to 32% in tests such as 3DMark Time Spy. This improvement is attributed to the increase in the number of CUDA cores and the adoption of GDDR7 memory, which provides greater bandwidth. This model has an eight-pin connector.
In addition to having a processor with support for 19 teraflops, this video card can also perform various AI activities, reaching up to 614 TOPS of performance without the need for an internet connection.
There is 8 GB of GDDR7 VRAM for running activities such as DLSS and local rendering, making it ideal for many heavier games. The long-awaited RTX 5060 GPU It arrives on the market with support for the main resolutions of the moment, ranging from Full HD (1920×1080) to the demanding 8K (7680×4320), passing through Quad HD (2560×1440) and the popular 4K (3840×2160).
In terms of connectivity, the board features three DisplayPort outputs and one HDMI, ensuring compatibility with a wide variety of monitors and offering advanced features for those looking for image quality and high performance on multiple screens.
GeForce RTX 5060 TI
More robust version of the intermediate line of NVIDIA, RTX 5060 Ti promises to deliver significant performance gains over RTX 4060 Ti. According to preliminary benchmarks, the new GPU presents a jump of up to 25% in gross performance, especially in tests like the 3DMark Time Spy and in games optimized for the Ada Lovelace Refresh architecture. The model has a sixteen-pin connector.
This advancement is driven by the 4.608 CUDA cores, new frequencies that reach 2.572 MHz in boost, and by the adoption of GDDR7 memory, which offers bandwidth of up to 448 GB / s — an improvement of more than 50% over the previous generation.
Available in versions with 8GB or 16GB of GDDR7 VRAM, RTX 5060 Ti maintains full support for DLSS 3.5, ray tracing and other artificial intelligence technologies, making it an ideal option for heavy gaming and creative applications.
The card is also compatible with multiple resolutions, supporting from Full HD (1920 × 1080) even the demanding 8K (7680 × 4320), through Quad HD (2560×1440) e Ultra HD / 4K (3840×2160). In terms of connectivity, the RTX 5060 Ti count with three DisplayPort outputs e an HDMI, allowing use on multiple monitors with high refresh rates.
Price and availability
Available for purchase from April 16, 2024, the new graphics cards come in two price points: Traditional RTX 5060 will be sold from $ 299 (approximately R$ 1.752 in direct conversion), while the RTX 5060 Ti comes to the market with a price that varies between 379 and 429 dollars (from R$ 2.227 to R$ 2.511 in direct conversion).
A NVIDIA also announced that the RTX 5060 for notebooks will be launched with models costing from US$ 1099 (R$ 6.420), with some already announced. They are:
- Acer Predator Helios Neo 16S;
- ASUS ROG Zephyrus G10;
- Dell Alienware 16 Area-51;
- Gigabyte AERO X16;
- HP Omen 16;
- Lenovo Legion 5i Pro;
- MSI Venture Pro 16 AI;
- Razer blade 16.
Some of the models were released at the beginning of the year, in CES 2025, so now they should be released with the new boards GeForce RTX 5060, as another option for the market. What did you think of the new video cards from NVIDIA? tell us Comment!
See also other features
With information: NVIDIA Newsroom.
Text review done by Jaime Ninece in 16 / 04 / 2025.
Discover more about Showmetech
Sign up to receive our latest news via email.
