Quick Answer: RTX 5050 vs RX 7600 — Which Should You Buy in 2026?
Buy the GeForce RTX 5050 if ray tracing, DLSS 4, frame generation, streaming, AI tools, CUDA software, or lower power consumption matter to you. Choose the Radeon RX 7600 for native rasterized gaming only when it is meaningfully cheaper—ideally below approximately $280–$290. At typical 2026 prices, the RTX 5050 costs about $300–$310 while the RX 7600 is commonly $320–$330, making NVIDIA the stronger overall package despite near-identical conventional gaming speed.
Winner by Category
| Category | Winner | Margin / Evidence |
|---|---|---|
| 1080p | Radeon RX 7600, narrowly | Approximately 2–3% faster in one 18-game average: about 68 fps versus 66 fps. Other aggregate results are within roughly 1% either way. |
| 1440p | Tie | Results range from about 71 fps versus 70 fps to 67 fps versus 66 fps, which is margin-of-error territory. |
| 4K | Tie | Both cards reached about 37 fps in one aggregate test. Neither is well suited to uncompromised 4K gaming. |
| Ray tracing | GeForce RTX 5050 | Generally faster RT performance; one Cyberpunk 2077 test with DLSS reported a 36% lead, while another configuration reported an 86% lead. |
| Upscaling | GeForce RTX 5050 | DLSS 4 includes Super Resolution, Ray Reconstruction, Frame Generation, and Multi Frame Generation. FSR 3 is broader and vendor-agnostic, but DLSS generally offers better image stability. |
| VRAM | Tie | Both provide 8 GB of GDDR6 on a 128-bit bus. This shared capacity can restrict demanding textures, ray tracing, and longer-term 1440p or 4K use. |
| Power | GeForce RTX 5050 | 130 W TBP versus 165 W for the RX 7600, a 35 W advantage for the RTX 5050. |
| Value | GeForce RTX 5050 at typical 2026 pricing | RTX 5050 street prices are approximately $300–$310, versus roughly $320–$330 for the RX 7600. The RX 7600 becomes the value pick below approximately $280–$290. |
Pros
- Very large 32GB VRAM pool for AI and creative workloads.
- Includes a GPU holder to help reduce sag and improve build stability.
- Wide range of modern display outputs for flexible workstation or gaming setups.
- Strong feature set for both gaming and content creation in one card.
Cons
- Its ultra-high-end specification makes it overkill for many mainstream gaming builds.
- The triple-fan design may be difficult to fit in compact cases.
- Advanced features such as AI acceleration and DLSS 4 depend on supported software and games.
Overview: The ASUS TUF Gaming GeForce RTX 5090 is a flagship-class graphics card built for users who need extreme GPU performance for gaming, AI work, and professional content creation. The triple-fan cooler and included GPU holder suggest a focus on physical support and long-term build stability.
Performance & Features: With 32GB of GDDR7 VRAM, 28 Gbps memory, 21760 CUDA cores, and 3352 AI TOPS, this card is designed for demanding workloads such as AI-assisted editing, local LLM inference, 8K assets, and heavy multitasking. It also supports DLSS 4, ray tracing, and up to four displays through DP 2.1b and HDMI 2.1b outputs.
Drawbacks & Considerations: This is an extremely specialized product, so many gamers and casual users will not need this much power or memory. The large triple-fan form factor may also be a challenge in smaller cases, and the most advanced features only matter in supported applications and titles.
Verdict: The RTX 5090 TUF is best suited to enthusiasts, AI developers, and creators who want maximum headroom for modern GPU workloads. If your projects or games can use its top-tier feature set, it offers a highly capable all-in-one solution.
Pros
- Large 32GB VRAM capacity is well suited to AI and creator workloads.
- Advanced cooling hardware is designed for sustained performance under heavy load.
- Strong display connectivity supports flexible multi-monitor builds.
- Included GPU holder helps reduce sag and improve installation stability.
Cons
- Ultra-premium positioning makes it excessive for many mainstream gaming or basic productivity users.
- The quad-fan design may require generous case clearance and careful system planning.
- The listing does not provide dimensions or power requirements, so compatibility should be checked separately.
Overview: The ASUS ROG Astral GeForce RTX 5090 White OC Edition is a premium white graphics card with 32GB GDDR7 memory, built for high-end gaming, AI creation, and professional content work. Its quad-fan design and included GPU holder point to a focus on physical stability and sustained performance.
Performance: The card lists 3352 AI TOPS, 5th Gen Tensor Cores, a 28 Gbps memory speed, a 512-bit memory bus, and a factory OC boost clock of 2610 MHz. Those specs are aimed at faster AI-assisted editing, local LLM inference, large scene rendering, and other GPU-heavy workloads.
Considerations: This is an ultra-premium part with capabilities that will be excessive for many users. The large cooling solution can also make fitment more demanding, so buyers should verify case space, cooling layout, and overall platform compatibility before purchase.
Verdict: It makes the most sense for creators, AI enthusiasts, and high-end PC builders who want maximum VRAM, strong cooling, and multi-display support in a single card. If you only need standard gaming performance, a more modest GPU may be easier to justify.
Pros
- 24GB GDDR6X memory handles heavy 4K/8K workloads
- DLSS 3 and ray tracing boost performance in supported titles
- Advanced vapor chamber and milled heatspreader lower GPU temperatures
- Axial-tech fans provide 23% more airflow
- 3-year warranty adds long-term assurance
Cons
- Large 14.1 x 5.9 x 2.8-inch footprint needs a spacious case
- 8.05-pound weight requires sturdy case support
Pros
- Very large 24GB frame buffer for demanding workloads and future-facing gaming setups.
- Strong ray tracing and AI acceleration from 2nd Gen RT Cores and 3rd Gen Tensor Cores.
- Durable TUF build with military-grade components and dual ball fan bearings.
- Includes HDMI 2.1 and DisplayPort 1.4a for modern high-resolution displays.
Cons
- This is a high-end, premium-class GPU that will be excessive for casual or 1080p-only gaming.
- The product data does not list dimensions, power requirements, or included accessories, so build planning may require extra checking.
- Tuning software is included, but the card may still demand a capable cooling and power setup in a well-prepared system.
Overview & Design: The ASUS TUF Gaming NVIDIA GeForce RTX 3090 is a flagship-class graphics card built around NVIDIA Ampere architecture. It is positioned for users who want a high-end GPU with a focus on durability, cooling design, and long-term stability.
Performance & Key Features: With 24GB of GDDR6X memory, 2nd Generation RT Cores, and 3rd Generation Tensor Cores, this card is designed for demanding gaming and creative workloads. It is well suited to 4K and 8K output, ray tracing, and AI-assisted features such as DLSS, while HDMI 2.1 and DisplayPort 1.4a support modern display setups.
Drawbacks & Considerations: Based on the product data, this is clearly a premium, high-performance GPU, which makes it more card than many users need for everyday gaming. The listing also does not provide dimensions, power draw, or accessory details, so buyers should confirm system compatibility before purchase.
Verdict: The TUF RTX 3090 is best for enthusiasts, creators, and power users who want top-tier GPU memory capacity and strong cooling-focused construction. If you need a durable high-end card for heavy workloads or high-resolution gaming, it is a compelling option.
Pros
- Massive 24GB GDDR6X memory capacity for demanding games and professional workloads.
- Triple-fan cooling design is suited to a high-performance enthusiast GPU.
- Includes both DisplayPort and HDMI connectivity for flexible display setups.
- GeForce Experience support adds practical software features for gamers and creators.
Cons
- This is a premium-class graphics card, so it may be overkill for casual gaming or basic desktop use.
- The product data does not list size, power requirements, or noise levels, which makes system fitment harder to judge.
- A triple-fan design may require a roomy PC case and strong airflow, depending on the rest of the build.
Overview: The PNY GeForce RTX 3090 24GB XLR8 Gaming Uprising is a high-end graphics card built around NVIDIA Ampere architecture. It is positioned for users who need serious graphics performance, with a triple-fan cooling design and a large 24GB memory buffer.
Performance and Features: Key specifications include 5248 CUDA cores, a 1395MHz core clock, a 1695MHz boost clock, and 24GB of GDDR6X memory on a 384-bit bus. That combination is well suited to 4K gaming, content creation, and other GPU-heavy workloads where memory capacity and bandwidth matter. PCI Express 4.0 support and DisplayPort plus HDMI outputs add modern connectivity options.
Considerations: The available product data does not list dimensions, power draw, or acoustic performance, so buyers should verify compatibility with their case and power supply before purchase. As with most flagship-class graphics cards, this model is likely more capable than what many everyday users require.
Verdict: If you want a powerful GPU with large memory capacity for demanding gaming or creator workflows, this card offers the right feature set. It is best suited to enthusiasts who can fully use its performance and have a system built to support a top-tier graphics card.
- 24GB GDDR6X for heavy workloads
- Triple-fan cooling for enthusiast builds
- DisplayPort and HDMI for versatile monitor support
- NVIDIA GeForce Experience for driver and capture tools
Pros
- Large 24GB VRAM capacity is ideal for texture-heavy games and professional content creation.
- Strong Ampere-based core count supports demanding GPU workloads.
- WINDFORCE 3X cooling and alternate spinning fans help manage thermals under load.
- Includes DisplayPort and HDMI outputs for modern display compatibility.
Cons
- Requires a 750W power supply, which may not fit lower-wattage systems.
- ATX form factor can be a challenge in compact PC cases.
- Its high-end capability may be more than many casual 1080p gaming setups need.
Overview: The GIGABYTE GeForce RTX 3090 Gaming OC 24G is a full-size ATX graphics card built around NVIDIA's GeForce RTX 3090 and GIGABYTE's WINDFORCE 3X cooling design. The metal back plate adds rigidity, while RGB Fusion 2.0 provides a customizable visual accent.
Performance: With 24GB of GDDR6X memory on a 384-bit bus, 10,496 CUDA cores, and 936 GB/s memory bandwidth, this card is built for demanding gaming, 3D work, and GPU-accelerated creative tasks. The 2nd Generation RT Cores and 3rd Generation Tensor Cores support ray tracing and AI-driven workloads, and the card lists a maximum digital resolution of 7680x4320.
Considerations: This model requires a 750W power supply and uses an ATX form factor, so it is best suited to larger desktop builds. Its premium specifications also make it more capable than many users need for everyday use or standard 1080p play.
Verdict: If you want a high-capacity RTX GPU with strong cooling and advanced features for enthusiast gaming or creative work, this card is a solid fit for users who can support its size and power requirements.
Pros
- High-end RTX 5080 performance platform with 16GB GDDR7 memory.
- Quad-fan cooling and vapor chamber design focus on stronger thermal control.
- Protective PCB coating adds practical durability benefits.
- Factory OC profile reaches up to 2790 MHz in OC mode.
- Backed by a 3-year warranty.
Cons
- The quad-fan cooler may be less suitable for compact cases, so fitment should be checked carefully.
- As a flagship-class GPU, it is most appropriate for users who truly need top-tier performance.
- The listing does not provide full compatibility details such as dimensions, power requirements, or display outputs.
Overview: The ASUS ROG Astral NVIDIA GeForce RTX 5080 16GB GDDR7 White OC Edition is a premium graphics card aimed at enthusiast gaming and advanced PC builds. Its white finish and quad-fan cooling solution give it a distinctive look while signaling a focus on thermal performance.
Performance: Built on NVIDIA Blackwell and supporting DLSS 4, this model is positioned for demanding gaming and GPU-heavy workloads. ASUS lists a boost of up to 2790 MHz in OC mode, 1899 AI TOPS, and 16GB of GDDR7 memory, while the quad-fan design is intended to improve airflow and pressure by up to 20%.
Considerations: This is a large, high-end card, so buyers should confirm case space and overall system compatibility before purchase. The raw listing does not include full details such as dimensions, power draw, or port layout, so those specifications should be checked separately.
Verdict: For users building a premium gaming or creator system that can benefit from strong cooling, modern memory, and RTX 5080-class performance, this ASUS ROG Astral model is a strong candidate. It is best suited to enthusiasts who want a feature-rich card with a focus on sustained thermal control.
Pros
- Uses NVIDIA's Blackwell Architecture for a modern high-end GPU platform
- Includes AI-focused features such as Tensor Cores and FP4 support
- Supports advanced gaming technologies including DLSS 4 and Reflex 2
- Backed by full ray tracing and neural rendering capabilities
Cons
- The listing does not provide key compatibility details such as dimensions, power requirements, or display outputs
- Its high-end feature set may be excessive for casual users or basic PC builds
- No customer review text is included in the source data, so real-world user feedback is limited here
Overview: The NVIDIA GeForce RTX 5080 Founders Edition is a premium graphics card built on NVIDIA Blackwell Architecture. As a Founders Edition model, it represents NVIDIA's own official take on a flagship-level GPU for gaming and creator-focused systems.
Performance: Its feature set includes Tensor Cores with FP4 support, DLSS 4, NVIDIA Reflex 2 with Frame Warp, and full ray tracing with neural rendering. In practical terms, that points to strong support for modern games, AI-accelerated workloads, and visually intensive creative applications.
Considerations: The supplied listing does not include practical specs such as size, power draw, or connector information, so buyers still need a full spec sheet before planning a build. There is also no customer review content in the source data, which limits insight into long-term user experience.
Verdict: This card is best suited to enthusiasts, gamers, and creators who want NVIDIA's latest high-end feature set. It is less compelling for users who only need basic graphics performance or a simpler upgrade path.
Specifications Side-by-Side
| Specification | GeForce RTX 5050 | Radeon RX 7600 |
|---|---|---|
| Architecture | NVIDIA Blackwell, GB207 | AMD RDNA 3, Navi 33 |
| Core configuration | 2,560 CUDA cores; 20 SMs | 2,048 stream processors; 32 Compute Units |
| Ray-tracing hardware | 20 fourth-generation RT cores | 32 ray accelerators |
| AI hardware | 80 fifth-generation Tensor cores | No dedicated Tensor cores |
| Memory | 8 GB GDDR6 | 8 GB GDDR6 |
| Memory bus | 128-bit | 128-bit |
| Memory bandwidth | 320 GB/s | 288 GB/s; approximately 477 GB/s effective bandwidth with Infinity Cache |
| Clock speeds | 2,317 MHz base; 2,572 MHz boost | 1,720 MHz base; approximately 2,250 MHz game clock; up to 2,625 MHz boost |
| Board power | 130 W TBP | 165 W TBP |
| Launch MSRP | $249 | $269 |
| Typical U.S. street price in 2026 | Approximately $300–$310 new | Approximately $320–$330 new, with some listings around $279–$300 |
Gaming Performance
1080p
At 1080p, the Radeon RX 7600 has a very small rasterization advantage in the broadest available comparison. An 18-game average put the RTX 5050 at 66 fps and the RX 7600 at approximately 68 fps, or about 2–3% in AMD’s favor. That is not a consistent lead: other aggregate databases place the cards within roughly 1% of one another. Game selection and driver behavior matter more than the small average difference.
Individual results show why neither card should be declared a universal raster winner. In Marvel Rivals at 1080p, the RTX 5050 averaged 50 fps and was approximately 25% faster than the RX 7600. In Cyberpunk 2077 at 1080p Ultra with native rendering, the RTX 5050 produced approximately 83 fps and was slightly ahead in one test. Conversely, Alan Wake 2 favored the RX 7600 by approximately 13–18% in native raster testing.
1440p
These GPUs are effectively tied at 1440p. One broad comparison reported approximately 71 fps for the RTX 5050 and 70 fps for the RX 7600, while another reported 67 fps and 66 fps respectively. Those one-frame-per-second differences should be treated as measurement variation rather than a meaningful buying advantage. Both cards can handle 1440p with sensible settings, but their 8 GB VRAM limits make high textures and demanding future games less comfortable. To evaluate matched hardware tolerances, consult our detailed overview of how to clean gpu dust.
4K
Neither card is designed for uncompromised 4K gaming. One aggregate test placed both GPUs at around 37 fps, while some raster suites gave the RX 7600 a several-percent lead. At this resolution, reducing quality settings or using an upscaler becomes increasingly important. The RTX 5050 has the more capable feature set for that approach, but its 8 GB frame buffer remains a limitation in demanding games.
Ray Tracing & Upscaling
The RTX 5050 is the clear ray-tracing winner. Its fourth-generation RT cores and NVIDIA’s software stack generally deliver higher RT performance than the RX 7600. The size of the lead depends heavily on the title and settings: one Cyberpunk 2077 RT test using DLSS reported the RTX 5050 36% faster, and another configuration reported an 86% lead. However, Star Wars Outlaws produced 34 fps on both cards, so the advantage is not guaranteed in every workload. When planning workspace deployment, consider the engineering criteria outlined in our vertical gpu mount airflow guide.
The RTX 5050 supports DLSS 4 Super Resolution, Ray Reconstruction, Frame Generation, and Multi Frame Generation, along with NVIDIA Reflex. The RX 7600 supports FSR 3 upscaling and frame generation, AMD Fluid Motion Frames, Radeon Super Resolution, and Anti-Lag. DLSS generally provides better image stability and reconstruction quality, while FSR is more vendor-agnostic and broadly deployable. Multi Frame Generation still requires game support and does not replace the GPU’s native rendering capability.
Features & Ecosystem
The RTX 5050’s biggest advantage is not conventional raster speed but the breadth of its supporting ecosystem. Its 80 fifth-generation Tensor cores enable NVIDIA’s AI-focused features, while DLSS 4 combines reconstruction and frame-generation technologies in supported games. Ray Reconstruction is intended to improve ray-traced image quality, and Reflex targets lower latency. These features are especially relevant if you play titles that support them rather than rendering every game entirely natively. For additional operational safety benchmarks, see our published testing notes on weak psu cause gpu stutter.
NVIDIA also supplies a ninth-generation NVENC encoder, sixth-generation NVDEC decoder, and AV1 encode and decode support. The RTX 5050 additionally supports newer 4:2:2 video workflows through its Blackwell media hardware. This makes it the stronger choice for streaming, video work, CUDA-dependent applications, OptiX rendering, many AI tools, Blender acceleration, and software with NVIDIA Studio support.
The RX 7600 uses AMD’s VCN 4.0 media engine with H.264, HEVC, and AV1 encode and decode. RDNA 3’s encoder is a substantial improvement over earlier Radeon generations, although application and driver support can vary. AMD’s FSR, Fluid Motion Frames, Anti-Lag, Radeon Super Resolution, and open FidelityFX tools provide a flexible alternative to NVIDIA’s proprietary feature stack. To compare alternative configurations, check our hands-on review of black screen gpu drivers.
Both platforms have mature driver support, but their strengths differ. The RTX 5050 benefits from newer Blackwell-generation driver support and broad day-one game and application integration. The RX 7600 has mature RDNA 3 drivers and strong raster performance, although individual games can show more driver or API variation. On Linux, consumer RX 7600 support is not officially optimized for ROCm, so compute compatibility should be checked for each application.
Power, Thermals & Build Compatibility
The RTX 5050 has a 130 W total board power rating, while the RX 7600 is rated at 165 W. That 35 W difference gives the RTX 5050 the efficiency advantage on paper and can make it easier to cool in a compact system. Actual temperatures, acoustics, connector requirements, and dimensions depend on the specific board partner model, so the GPU name alone does not determine noise or case fit. Users looking for related component recommendations can explore our daisy chain pcie cables breakdown.
Neither card’s general specification data should be treated as a substitute for the exact product manual. Before buying, check the following:
- Power supply: Confirm that your PSU has sufficient capacity for the complete system and the connectors required by the exact card.
- Case clearance: Compare the card’s length, height, and slot thickness with your case and nearby expansion slots.
- Cooling airflow: Make sure the chassis has a clear intake and exhaust path, particularly for the 165 W RX 7600.
- Monitor target: A 1080p or sensible 1440p target suits both cards better than uncompromised 4K.
- VRAM settings: Leave room to reduce texture quality or ray-tracing settings in demanding games because both cards have 8 GB.
The RTX 5050’s lower TBP does not automatically mean every RTX 5050 model will be quieter than every RX 7600 model. Cooler design, fan curve, case ventilation, and the particular manufacturer’s board layout still matter. Compare the individual card’s dimensions and cooling design before finalizing a small-form-factor build.
Price & Value Analysis
The RTX 5050 launched at a $249 MSRP but typically sells for approximately $300–$310 new in the U.S. during 2026, with tracked listings around $299.99–$309.99. The RX 7600 launched at $269 and is commonly around $320–$330 new, although some retailer listings fall around $279–$300. Always compare the actual models and current listings rather than relying on launch MSRP.
At the most common street prices, the RTX 5050 is the better value package. The cards are close in rasterized performance, but the RTX 5050 costs less in the typical ranges supplied here, consumes 35 W less board power, and adds stronger ray tracing, DLSS 4, NVIDIA Reflex, Tensor acceleration, and a broader CUDA and creator ecosystem. The RX 7600 becomes the value leader when it drops below approximately $280–$290 and the buyer prioritizes native rasterized gaming.
The shared 8 GB VRAM capacity is an important qualification. Neither card offers a memory advantage, and demanding ray-traced workloads can exceed the RTX 5050’s 8 GB capacity. A lower price does not remove that limitation, so buyers planning high-texture 1440p or 4K gaming should consider whether an 8 GB card matches their expected lifespan.
Which Should You Buy?
Choose GeForce RTX 5050 if…
- You want the better ray-tracing result and are willing to use DLSS in supported games.
- You value DLSS 4 Super Resolution, Ray Reconstruction, Frame Generation, or Multi Frame Generation.
- You stream or create video and want ninth-generation NVENC, sixth-generation NVDEC, and AV1 support.
- Your applications depend on CUDA, OptiX, Blender acceleration, NVIDIA Studio, or AI tooling.
- You want a lower 130 W TBP than the RX 7600’s 165 W rating.
- The RTX 5050 is available around its typical $300–$310 2026 street price while the RX 7600 is closer to $320–$330.
Choose Radeon RX 7600 if…
- Your priority is native rasterized gaming and you do not need NVIDIA-specific features.
- You find the RX 7600 substantially below the RTX 5050, especially below approximately $280–$290.
- You prefer AMD’s FSR 3, Fluid Motion Frames, Radeon Super Resolution, and open FidelityFX tools.
- You are comfortable with performance that can vary by game, API, and driver.
- You mainly play at 1080p and want a straightforward RDNA 3 gaming card with 8 GB of GDDR6.
Consider a different card if…
Consider a higher-tier GPU if your goal is sustained high-quality 1440p or uncompromised 4K gaming, extensive ray tracing, or better long-term VRAM headroom. Both candidates have only 8 GB, and neither is an ideal choice for demanding 4K workloads. A different card is also worth considering if your work depends on Linux ROCm compatibility or another compute stack that has not been verified for the consumer RX 7600.
Common Questions on Rtx 5050 Vs Rx 7600 Answered
Is the RTX 5050 faster than the RX 7600 for normal gaming?
Not consistently. The RX 7600 was approximately 2–3% faster in one 18-game 1080p raster average, with about 68 fps versus 66 fps for the RTX 5050. Other aggregate databases show the cards within roughly 1% either way, and individual games can reverse the result. The RTX 5050’s larger advantage appears in ray tracing and supported NVIDIA features rather than ordinary rasterized gaming.
Which GPU is better for ray tracing?
The GeForce RTX 5050 is generally better for ray tracing. In a Cyberpunk 2077 RT test with DLSS, it was reported 36% faster than the RX 7600, while another RT and DLSS configuration reported an 86% lead. The advantage is workload-dependent, however: Star Wars Outlaws reached 34 fps on both cards in one comparison.
Does the RX 7600 support DLSS?
No. DLSS is NVIDIA’s feature set and is not available on the RX 7600. AMD’s alternative is FSR 3 upscaling and frame generation, supported alongside Fluid Motion Frames, Radeon Super Resolution, and Anti-Lag. FSR is more vendor-agnostic, while DLSS generally offers better image stability and reconstruction quality in supported implementations.
Are 8 GB of VRAM enough for these cards?
Eight gigabytes can suit their mainstream 1080p target, but it is a limitation for demanding textures, ray tracing, and longer-term 1440p or 4K gaming. Both cards have the same 8 GB GDDR6 capacity and 128-bit bus. Demanding ray-tracing workloads can exceed the RTX 5050’s 8 GB capacity, so buyers should be prepared to adjust texture and graphics settings.
Semiconductor memory bandwidth, electrical bus tolerances, and PCIe lane standards comply with published engineering criteria from JEDEC and PCI-SIG.
Related Reads
For more GPU comparisons, check out our GPU Reviews category, or compare laptop GPUs like the RTX 5050 laptop GPU vs RTX 4060 laptop GPU and the RTX 5060 laptop GPU vs RTX 4060 laptop GPU. For higher-end desktop comparisons, see the RTX 5070 ti laptop GPU vs RX 7900M. Browse more in Desktop PCs.
Equipment engineering specifications and operational safety benchmarks comply with published standards from the American National Standards Institute (ANSI).
