TurboVS
Manli Stellar GeForce RTX 5070 OC specifications and in-depth review

Manli Stellar GeForce RTX 5070 OC

Manufacturer: Manli

The Manli Stellar GeForce RTX 5070 OC is a desktop graphics card based on NVIDIA's Blackwell architecture, manufactured on a 5 nm process with 31,100 million transistors. It runs at a base clock of 2325 MHz, reaching a boost speed of 2542 MHz, and delivers 31.24 TFLOPS of floating-point performance alongside a pixel rate of 203.4 GPixel/s. The card includes RGB lighting and supports up to four displays simultaneously.

On the memory side, the card is equipped with 12GB of GDDR7 VRAM across a 192-bit bus, with an effective speed of 28,000 MHz and a maximum bandwidth of 672 GB/s. It supports DirectX 12 Ultimate, OpenGL 4.6, OpenCL 3, ray tracing, DLSS, and Intel Resizable BAR. Connectivity options include one HDMI 2.1b port and three DisplayPort outputs, with a TDP of 250W and physical dimensions of 312 mm by 127 mm. The card connects via PCIe 5 and supports ECC memory.

Pros
  • Supports up to four displays simultaneously, making it suitable for multi-monitor setups
  • GDDR7 memory with a 672 GB/s bandwidth ceiling and an effective speed of 28,000 MHz enables fast data throughput for demanding workloads
  • Ray tracing and DLSS support are both present, adding flexibility for compatible rendering pipelines
  • ECC memory support adds a degree of data reliability for compute-oriented tasks
  • Intel Resizable BAR is included, allowing the CPU broader access to GPU memory during operation
  • RGB lighting is built in, offering visual customization without requiring additional hardware
Cons
  • The 192-bit memory bus is narrower than what some cards in this category offer, which can constrain memory throughput under heavy loads
  • No USB-C output is available, limiting compatibility with certain modern displays and adapters
  • Air-water cooling is not supported, so users seeking liquid cooling integration will need to look elsewhere
  • With a TDP of 250W, the card places a meaningful demand on system power delivery
  • The single HDMI port may be limiting for users who rely on multiple HDMI-connected devices
Who is this for?

This card is well matched to users who run demanding rendering and compute workloads, given its 31.24 TFLOPS of floating-point performance, ECC memory support, and Double Precision Floating Point capability. Those working with multi-display setups of up to four screens will find the output configuration practical, while the inclusion of ray tracing, DLSS, and Intel Resizable BAR makes it a reasonable fit for users running compatible graphics pipelines that benefit from these features. The 12GB of GDDR7 VRAM with 672 GB/s bandwidth also suits scenarios where large datasets or textures need to be moved quickly through memory.

Who is this NOT for?

Users who require liquid cooling integration will find this card unsuitable, as air-water cooling is not supported. Those with a single HDMI display chain who rely on USB-C connectivity for modern monitors or docking setups will also find the port selection limiting, since no USB-C output is available. Additionally, systems with constrained power delivery may struggle to accommodate the card's 250W TDP, making it a poor fit for compact or low-wattage builds where power efficiency is a priority.

Performance:

GPU clock speed 2325 MHz
GPU turbo 2542 MHz
pixel rate 203.4 GPixel/s
floating-point performance 31.24 TFLOPS
texture rate 488.1 GTexels/s
GPU memory speed 1750 MHz
shading units 6144
texture mapping units (TMUs) 192
render output units (ROPs) 80
Has Double Precision Floating Point (DPFP) ✓

The Performance section of this card centers on a base GPU clock of 2325 MHz that climbs to a boost speed of 2542 MHz, supporting a floating-point output of 31.24 TFLOPS. Rasterization throughput is handled by 80 render output units delivering a pixel rate of 203.4 GPixel/s, while 192 texture mapping units achieve a texture rate of 488.1 GTexels/s. The GPU includes 6,144 shading units and operates its memory at 1750 MHz, and it also supports Double Precision Floating Point (DPFP) for workloads that require high numerical accuracy.

Memory:

effective memory speed 28000 MHz
maximum memory bandwidth 672 GB/s
VRAM 12GB
GDDR version GDDR7
memory bus width 192-bit
Supports ECC memory ✓

This card is equipped with 12GB of GDDR7 VRAM running across a 192-bit memory bus, reaching an effective speed of 28,000 MHz and a maximum bandwidth of 672 GB/s. ECC memory support is also included, adding a layer of data integrity protection for workloads that require it.

Features:

DirectX version DirectX 12 Ultimate
OpenGL version 4.6
OpenCL version 3
Supports multi-display technology ✓
supports ray tracing ✓
Supports 3D ✓
supports DLSS ✓
has XeSS (XMX) ✕
AMD SAM / Intel Resizable BAR Intel Resizable BAR
has LHR ✕
has RGB lighting ✓
supported displays 4

The card supports DirectX 12 Ultimate, OpenGL 4.6, and OpenCL 3, covering a broad range of graphics and compute workloads. Ray tracing, DLSS, and stereoscopic 3D are all supported, while XeSS (XMX) and LHR are not present. Multi-display technology is enabled with support for up to four screens simultaneously, and Intel Resizable BAR is included to help streamline data transfers between the CPU and GPU. The card also features RGB lighting for those who want visual customization in their build.

Ports:

has an HDMI output ✓
HDMI ports 1
HDMI version HDMI 2.1b
DisplayPort outputs 3
USB-C ports 0
DVI outputs 0
mini DisplayPort outputs 0

The card's output configuration consists of one HDMI 2.1b port and three DisplayPort outputs, giving a total of four display connections. There are no USB-C, DVI, or mini DisplayPort outputs on this model.

General info:

GPU architecture Blackwell
Thermal Design Power (TDP) 250W
PCI Express (PCIe) version 5
semiconductor size 5 nm
number of transistors 31100 million
Has air-water cooling ✕
width 312 mm
height 127 mm

Built on the Blackwell architecture and fabricated using a 5 nm process, this card packs 31,100 million transistors onto its die. It connects via PCIe 5 and carries a Thermal Design Power of 250W. The card measures 312 mm in width and 127 mm in height, and it does not include air-water cooling. A standard air cooling solution is used instead.

Final Verdict

The Manli Stellar GeForce RTX 5070 OC is a capable graphics card built around the Blackwell architecture, bringing together a solid set of specifications that serve both graphics-intensive and compute-oriented workloads. Its 12GB of GDDR7 VRAM paired with 672 GB/s of memory bandwidth stands out as one of its most defining attributes, complemented by ray tracing, DLSS, ECC memory support, and a four-display output configuration. The card does carry certain constraints — most notably its lack of USB-C output, absence of air-water cooling support, and a 250W power requirement that demands a well-specced system. Taken as a whole, it represents a well-rounded option for users who can accommodate its power and thermal requirements and whose workflows align with the features it genuinely delivers.