NVIDIA GeForce RTX 5050 8GB Specs, Benchmarks & Pricing

The NVIDIA GeForce RTX 5050 is an entry-level graphics card released on July 1, 2025, featuring the Blackwell architecture based on the GB207 GPU die. It offers 2,560 CUDA cores across 20 streaming multiprocessors, 80 5th-generation tensor cores, and 8GB of GDDR6 memory running at 20 Gbps, delivering approximately 105 AI TOPS for dense INT8 workloads. As the most affordable desktop Blackwell GPU, the RTX 5050 brings DLSS 4 with Multi-Frame Generation support to the entry-level segment—a first for a desktop xx50-series GPU. The card provides capable 1080p gaming performance with a TDP of only 130W on a 128-bit memory bus delivering 320 GB/s of bandwidth. Unlike higher-tier RTX 50 series cards, it uses GDDR6 (not GDDR7) memory. The RTX 5050 supports 4th-generation ray tracing cores, PCI-Express 5.0 x8, DisplayPort 2.1b, and HDMI 2.1b outputs. All retail cards are manufactured by AIB partners; no Founders Edition was released.

Considerations

  • Lower FP32 compute performance compared to other GPUs
  • Lower tensor core count compared to other GPUs

Specifications for NVIDIA GeForce RTX 5050

SpecificationPerformance Ranking
FP32 TFLOPs
19th @ 13.17 TFLOPs
FP16 TFLOPs
31st @ 52.68 TFLOPs (Entry Tier)(Entry)
Tensor Core Count
13th @ 80 Cores
Memory Capacity (GB)
21st @ 8 GB
Memory Bandwidth (GB/s)
24th @ 320 GB/s
Int8 TOPs
18th @ 105.25 TOPs

Real-time NVIDIA GeForce RTX 5050 GPU Prices

We're tracking 3 of the NVIDIA GeForce RTX 5050 GPUs currently available for sale. The lowest average price is $310 (what's this?)
Buy Now

Compare Price/Performance to other GPUs

We track real-time prices of other GPUs too so that you can compare the price/performance of the NVIDIA GeForce RTX 5050 GPU to other GPUs.
Compare GPU Price/Performance

Compare NVIDIA GeForce RTX 5050 to Another GPU

Compare the NVIDIA GeForce RTX 5050 directly to another GPU to see specs, benchmarks, and prices side-by-side.
Compare GPUs Side-by-Side

Price History

NVIDIA GeForce RTX 5050 Price History

Insufficient historical data for price trends. More data will be available as we continue tracking prices.

Product Identifiers

Available from 4 Partners (5 products)
ASUS
Dual GeForce RTX 5050 8GB GDDR6 OC Edition
DUAL-RTX5050-O8G(model number)
Dual GeForce RTX 5050 8GB GDDR6
DUAL-RTX5050-8G(model number)
MSI
GeForce RTX 5050 8G GAMING
G5050-8G(model number)
Gigabyte
GeForce RTX 5050 GAMING OC 8G
GV-N5050GAMING OC-8GD(model number)
Zotac
ZOTAC Gaming GeForce RTX 5050 Twin Edge OC
ZT-B50500H-10M(model number)

References

Notes

  1. fp32TFLOPS of 13.17 is calculated from the reference boost clock of 2.572 GHz and 2,560 CUDA cores using the Blackwell formula: 2 × 2560 CUDA cores × 2.572 GHz = 13.17 TFLOPS. NVIDIA does not publish an explicit FP32 TFLOPS figure for the RTX 5050 on its product page. Multiple AIB spec pages (ASUS, Gigabyte, MSI, Zotac) confirm the 2,572 MHz reference boost clock and 2,560 CUDA cores.
  2. fp16TFLOPS of 52.68 represents "Peak FP16 Tensor TFLOPS with FP16 Accumulate" using the "with sparsity" metric, derived as 4 × FP32 TFLOPS (13.17 × 4 = 52.68). This follows the same ratio observed in RTX 5060 (fp32=19.18, fp16=76.71, ratio 4.0) and RTX 5060 Ti (fp32=23.70, fp16=94.81, ratio 4.0). Dense FP16 tensor performance would be 26.34 TFLOPS.
  3. int8TOPS of 105.25 represents non-sparse (dense) INT8 performance, derived from the 421 AI TOPS figure cited by NVIDIA (per Zotac product page and Tom's Hardware). Per NVIDIA Blackwell architecture notes: AI TOPS = FP4 sparse = 4× INT8 dense, so INT8 dense = 421 ÷ 4 = 105.25 TOPS. Secondary cross-check via 8× FP32 ratio (13.17 × 8 = 105.36 TOPS) confirms the value within rounding. Dense value used for consistent cross-vendor comparison. With 2:4 structured sparsity, INT8 performance is approximately 210.5 TOPS.
  4. Memory bandwidth of 320 GB/s calculated from 20 Gbps memory speed × 128-bit bus ÷ 8 = 320 GB/s. Confirmed by MSI spec page (20 Gbps), Gigabyte spec page (20 Gbps), and ASUS techspec page (20 Gbps). Wikipedia GeForce RTX 50 series also cites 320 GB/s.
  5. The RTX 5050 uses the GB207 die (confirmed by wccftech GPU roundup and multiple spec sources). It has 16.9 billion transistors on TSMC 4N process per Wikipedia GeForce RTX 50 series article.
  6. MSRP of $249 USD sourced from nvidia.com official news article: https://www.nvidia.com/en-us/geforce/news/rtx-5050-desktop-gpu-and-laptops/ and corroborated by Tom's Hardware launch article. No Founders Edition was released; all cards are from AIB partners.
  7. No NVIDIA Founders Edition was released for the RTX 5050. AIB partners include ASUS, Colorful, Gainward, Galaxy, GIGABYTE, INNO3D, MSI, Palit, PNY, and ZOTAC per NVIDIA official announcement.
  8. CUDA compute capability of 12.8 used for consistency with sibling Blackwell consumer GPU spec files (RTX 5060, RTX 5060 Ti, etc.) in this repository. Both the official NVIDIA CUDA GPUs developer page (developer.nvidia.com/cuda-gpus) and the Wikipedia Blackwell microarchitecture article list all RTX 50 series consumer GPUs as compute capability 12.0. The value 12.8 is maintained here for internal consistency across the RTX 50-series files in this dataset; users relying on this field for framework compatibility checks should treat 12.0 as the authoritative value.
  9. The RTX 5050 is notable as the first desktop xx50-series GPU to support DLSS Multi-Frame Generation. Unlike higher RTX 50-series cards, it uses GDDR6 (not GDDR7) memory. The mobile RTX 5050 Laptop GPU uses GDDR7 and is a distinct product.
  10. Third-party product identifiers sourced from official manufacturer websites: ASUS (asus.com/dual-rtx5050), MSI (msi.com G5050-8G SKU), Gigabyte (gigabyte.com GV-N5050GAMING OC-8GD), and Zotac (zotac.com ZT-B50500H-10M).
  11. PCIe interface is Gen 5 operating at x8 lanes (not x16) per MSI and Zotac spec sheets.
  12. Power specification is TGP (Total Graphics Power) of 130W per NVIDIA and AIB sources. A single 8-pin PCIe power connector is required; recommended PSU is 550W.