How to Check What GPU You Have (Windows 10 & 11) | 2026 Guide

To check what GPU you have on Windows, press Ctrl + Shift + Esc to open Task Manager, click the Performance tab, then click GPU in the left sidebar – your graphics card name shows up in the top-right corner. The whole thing takes about 10 seconds and works on both Windows 10 and Windows 11 without installing anything.

But Task Manager only tells you the marketing name. Sometimes you need deeper info – how much VRAM you have, whether your drivers are current, or if your laptop is running games on the integrated chip instead of the dedicated GPU. This guide covers eight different methods to identify your graphics card, explains what the specs mean, and walks through the most common problems people run into.

GamerHardware.org is reader-supported. When you buy through links on our site, we may earn an affiliate commission at no extra cost to you. Learn more.

8 Ways to Check Your GPU (Visual Overview)

Every method below is ranked by speed and detail. Most people only need Method 1. The rest exist for specific situations – driver troubleshooting, remote support, scripting, support tickets, or hardware verification before selling a card.

At a glance – which tool gives you which data:

Method 1: Task Manager (Fastest – 10 Seconds)

Task Manager is the go-to method because it is fast, accurate, and already installed on every Windows PC.

1. Press Ctrl + Shift + Esc to open Task Manager directly. On Windows 11, you can also right-click the taskbar and pick “Task Manager.”
2. Click the Performance tab along the top. If you see a simplified view with just running apps, click “More details” at the bottom-left first.
3. Click GPU in the left sidebar. Your GPU name appears in the top-right corner – something like “NVIDIA GeForce RTX 4070 Ti” or “AMD Radeon RX 7900 XTX.”

The Performance panel also shows live data: GPU temperature, GPU utilization, dedicated memory usage, and shared memory usage. If you see GPU 0 and GPU 1, your system has both an integrated and a dedicated graphics card. GPU 0 is usually the integrated chip. GPU 1 is the dedicated card you want to game on.

Limitation: Task Manager shows the commercial product name, not the silicon-level chip name. It won’t tell you the difference between a laptop RTX 4060 and a desktop RTX 4060. For that, you need GPU-Z.

Method 2: Windows Settings (Mouse-Only)

If keyboard shortcuts aren’t your thing, Settings works with just the mouse. This approach is also handy on tablets running Windows.

1. Open Settings by pressing Windows + I, or click the Start menu and select the gear icon.
2. Go to System, then Display.
3. Scroll down and click “Advanced display” (Windows 11) or “Advanced display settings” (Windows 10).
4. Your GPU model appears under “Display information” next to the monitor name and refresh rate.

On multi-monitor setups, you can switch between displays at the top of this page to see which GPU drives each monitor. Handy if you have one plugged into the motherboard and another into the dedicated card.

Limitation: This only shows the GPU connected to the display. If you have a second GPU not connected to any monitor (rare, but it happens with compute cards), it won’t appear here.

Windows 10 note: the Advanced display setting is labeled “Advanced display settings” under Display instead of “Advanced display.”

Method 3: dxdiag (Most Detail Without Downloads)

dxdiag has been in Windows since the DirectX era of the late 1990s. It packs more GPU info into one screen than any other built-in tool.

1. Press Windows + R to open the Run dialog.
2. Type dxdiag and press Enter.
3. Click the Display tab. If you have multiple GPUs, there may be a “Display 1” and “Display 2” tab, plus a “Render” tab.

The Display tab shows your GPU name, manufacturer, chip type, DAC type, total memory, display mode, and your current driver version with its date. The “Save All Information” button at the bottom exports everything to a text file – useful when someone helping you remotely asks “what GPU do you have?”

Watch out: The “Approx. Total Memory” field is misleading. It adds dedicated VRAM and shared system memory together. An 8GB GPU might show 16GB here. For accurate VRAM, use Task Manager’s “Dedicated GPU Memory” field or check with GPU-Z.

Method 4: PowerShell One-Liner (Scripting and Remote Admin)

PowerShell pulls GPU info without any GUI. This is the fastest method if you already have a terminal open, and you can run it on remote machines.

This returns every GPU in the system – integrated and dedicated – along with VRAM in bytes and driver version strings. To convert VRAM bytes to GB, divide by 1073741824. Or run this version instead:

This outputs something like:

Note: The Win32_VideoController class has a known issue where AdapterRAM caps at 4GB for cards with more than 4GB of VRAM. This is a WMI limitation, not a PowerShell bug. If you see “4.0” for a card you know has 8GB or 12GB, the VRAM field is truncated. Use GPU-Z or Task Manager for accurate VRAM on high-end cards.

For Command Prompt instead of PowerShell, the equivalent command is:

On Windows 7 and 8 where Get-CimInstance isn’t available, substitute Get-WmiObject Win32_VideoController | Select Name, AdapterRAM, DriverVersion – it reads the same WMI class.

Windows 10/11 note: both ship PowerShell 5.1 and support Get-CimInstance. If you see “Get-CimInstance is not recognized”, open Windows PowerShell instead of Command Prompt.

Method 5: Device Manager (For Driver Problems)

Device Manager is where you go when your GPU is not working correctly. It shows driver status, error codes, and lets you roll back or update drivers.

1. Right-click the Start button and select Device Manager. (Keyboard shortcut: Windows + X, then M.)
2. Expand “Display adapters.” Your GPU model appears here.
3. Right-click the GPU for options: Update driver, Disable device, Uninstall device, or Properties for driver details.

If you see a yellow triangle with an exclamation mark on the GPU entry, Windows detected a driver error. Double-click the device and check the “Device status” field for an error code (like Code 43, which means Windows stopped the device because it reported problems).

If you see “Microsoft Basic Display Adapter” instead of an NVIDIA or AMD card, your GPU drivers are missing or corrupted. Windows fell back to a generic software renderer. The fix: download the correct driver from your GPU manufacturer’s website.

Method 6: GPU-Z (Power Users)

GPU-Z from TechPowerUp is a free, portable tool (no install needed – just run the .exe) that shows every technical detail about your graphics card.

What GPU-Z gives you that built-in tools don’t:

• Exact chip name – like AD104-251 or Navi 32, not just “RTX 4070”
• Memory bus width – 128-bit, 192-bit, 256-bit (affects bandwidth and performance)
• Shader count – CUDA cores (NVIDIA) or stream processors (AMD)
• BIOS version – useful for custom BIOS flashing and verifying factory settings
• Fake GPU detection – compares reported specs against known databases to catch counterfeits
• Sensor tab – real-time clock speeds, temperatures, fan RPM, power draw, and voltage

GPU-Z is the gold standard if you’re buying a used card. Scammers flash cheap GPUs with fake BIOS data to make an old GTX 960 show up as an RTX 3060 in Device Manager and dxdiag. GPU-Z catches this by verifying the silicon-level chip ID against its database.

Download it from techpowerup.com/gpuz. The file is under 10MB and runs without installation.

Other free third-party options include HWiNFO64 (more detailed hardware sensors than GPU-Z, also portable) and Speccy from Piriform (simpler UI, less granular). GPU-Z stays the default choice for counterfeit detection because of its PCI device-ID database.

Method 7: System Information (msinfo32)

System Information has been built into Windows since Windows 7 and shows adapter data in a hierarchical tree. It’s the right tool when a support tech asks for a full system snapshot, not just a GPU name.

1. Press Windows + R to open the Run dialog.
2. Type msinfo32 and press Enter.
3. Expand “Components” in the left pane, then click “Display.”

The right pane lists Name, Adapter Type, Adapter Description, Adapter RAM, Adapter DAC Type, Installed Drivers, Driver Version, INF File, Current Mode, Monitor, and more. Each display adapter in the system gets its own block – useful when you have an integrated GPU and a dedicated card side by side.

One honest limitation: the “Adapter RAM” field has the same 4 GB cap as the WMI Win32_VideoController class, so on a card with 8 GB or 12 GB of VRAM it will still read around 4 GB. For accurate VRAM, cross-check with Method 1 (Task Manager) or the dedicated VRAM check section.

Export tip: File > Export saves the entire report as a .txt file. Useful when a support tech asks for a system snapshot covering BIOS, installed drivers, running services, and hardware resources in one attachment.

When to use this vs dxdiag: msinfo32 shows more context (installed drivers list, BIOS, startup programs), while dxdiag shows DirectX feature levels and Direct3D acceleration status.

Method 8: Vendor Tools (NVIDIA, AMD, Intel)

These apps are already on your PC if you installed a driver. They show per-vendor data that built-in tools can’t, like CUDA core counts, current core and memory clocks, VBIOS version, and Resizable BAR status.

NVIDIA Control Panel

Right-click the desktop and pick NVIDIA Control Panel. Click the System Information link in the bottom-left corner. The Components and Display tabs show driver version, video memory, BIOS version, CUDA cores, DirectX version, bus interface (PCIe 4.0 x16, for example), device ID, and part number. This is the place to verify an exact OEM board or catch BIOS mismatches after a flash.

AMD Software: Adrenalin Edition

Right-click the desktop and pick AMD Software. Go to the Performance tab, then Tuning, then expand the System pane. You’ll see GPU model, driver and package version, VBIOS version, and Windows edition. The Software tab in Adrenalin also lists detected display and audio driver versions – handy when a game refuses to launch because of an old audio driver.

Intel Graphics Command Center and Intel Arc Control

Search Start for Intel Graphics Command Center (integrated Intel graphics like UHD or Iris Xe) or Intel Arc Control (Arc A-series dedicated cards). The Home or System tab shows GPU name, driver version, available VRAM, and supported feature levels. Arc Control also exposes Resizable BAR status, which is a major factor in Arc gaming performance.

Honest caveat: these apps only work if the matching vendor driver is installed. A fresh Windows install without GPU drivers won’t show any vendor tool – use msinfo32 or dxdiag in that case.

How to Check Your VRAM

VRAM (Video RAM) determines how many textures, frames, and assets your GPU can hold at once. Running out of VRAM causes stuttering, texture pop-in, and crashes. Here’s where to find accurate VRAM numbers – and which tools lie about it.

Here’s the quick version: open Task Manager, go to Performance, click GPU, and look at “Dedicated GPU Memory” at the bottom. That number is your actual VRAM. Ignore “Shared GPU Memory” – that’s system RAM that Windows can lend to the GPU in emergencies, and games don’t count it as usable VRAM.

Why VRAM matters for gaming: modern games at 1440p typically need 8GB of VRAM. At 4K with high textures, 10-12GB is a comfortable minimum. Games like Hogwarts Legacy and Cyberpunk 2077 can use over 10GB at max settings in 4K. If your VRAM isn’t enough, the game drops textures to lower resolution mid-play, causing visible pop-in.

Integrated vs. Dedicated GPU – How to Tell

An integrated GPU lives inside the CPU die. Intel calls theirs UHD Graphics or Iris Xe. AMD’s integrated chips go by Radeon Graphics or Radeon 780M. A dedicated GPU is a separate card in the PCIe slot – the RTX 4070, RX 7800 XT, and Arc A770 are dedicated cards.

The performance gap is massive. Intel UHD 770 pushes around 30 FPS in Valorant at 1080p low settings. An RTX 4070 pushes 300+ FPS in the same scenario. That’s a 10x difference.

How to tell which one your system has (or if it has both):

1. Open Task Manager (Ctrl + Shift + Esc) and click Performance.
2. Count the GPU entries in the sidebar. One entry means one GPU. Two entries means one integrated + one dedicated.
3. GPU 0 is typically the integrated chip. GPU 1 is the dedicated card.

If only one GPU appears and it says “Intel UHD” or “AMD Radeon Graphics” (without a specific model number like RX 7800), you only have an integrated GPU. Gaming on integrated graphics is possible for lighter titles, but you’ll need a dedicated card for anything demanding.

Laptop GPU Detection (Why It Matters)

Laptops add a twist: most gaming laptops have both an integrated and dedicated GPU, and the system switches between them automatically. This is called “Optimus” on NVIDIA laptops and “SmartShift” on AMD laptops. The problem? Sometimes games run on the integrated chip by accident, and you get 20 FPS instead of 120.

To check which GPU a game is actually using:

1. Open Task Manager while the game is running.
2. Click the Performance tab and check both GPU 0 and GPU 1.
3. The one with high utilization (60-100%) is the active GPU. The other should sit near 0-5%.

If you see the integrated GPU maxed out while the dedicated card is idle, force the game to use the right GPU. On Windows 11: Settings > System > Display > Graphics > select the game’s .exe > Options > set to “High performance.” On NVIDIA, open NVIDIA Control Panel > Manage 3D Settings > Program Settings > add the game > set to “High-performance NVIDIA processor.”

Another laptop-specific issue: a laptop “RTX 4060” is not the same as a desktop RTX 4060. Laptop variants run at lower power limits (anywhere from 60W to 140W vs. 115W fixed on desktop) and often have fewer CUDA cores. GPU-Z shows the exact wattage and chip variant if you need to compare.

Two laptops with the same GPU badge can perform very differently because of Total Graphics Power limits the OEM chose. A 140W RTX 4060 Laptop in a thick gaming chassis outperforms an 80W RTX 4060 Laptop in a thin notebook by 20-30%. Here’s roughly what to expect:

GPU-Z’s “Default Board Power” field shows the configured TGP for your specific unit.

GPU Naming Decoded: NVIDIA, AMD, and Intel

GPU model names look like alphanumeric soup until you learn the pattern. Here’s how to read them.

NVIDIA Naming

Format: RTX [Generation][Tier] [Suffix]

• Generation: First two digits. 50 = RTX 5000 series (2025), 40 = RTX 4000 (2022-2024), 30 = RTX 3000 (2020-2022).
• Tier: Last two digits. 90 = flagship, 80 = high-end, 70 = mid-range, 60 = mainstream, 50 = entry.
• Ti: More CUDA cores and higher clock speeds than the base model. Usually 10-15% faster.
• Super: Mid-cycle refresh. Slightly more cores and/or faster memory than the original.
• RTX vs. GTX: RTX supports ray tracing and DLSS. GTX (1000/1600 series) doesn’t. GTX cards still work fine – they just lack those two features.

So an RTX 5070 Ti Super is a mid-cycle-refreshed, enhanced-tier card from the 5000 generation. Not confusing at all.

AMD Naming

Format: RX [Generation][Tier]0 [Suffix]

• Generation: First digit. 9 = RX 9000 (2025), 7 = RX 7000 (2022-2024), 6 = RX 6000 (2020-2022).
• Tier: Second digit. 9 = flagship, 8 = high-end, 7 = mid-range, 6 = mainstream.
• XT: Faster variant with more stream processors and/or higher clocks. Similar to NVIDIA’s Ti.
• XTX: The top SKU within a tier. Only used on select flagship models.

Intel Arc Naming

Format: Arc A[Tier][Variant]

• A7xx: High-end (Arc A770, A750). Competitive with RTX 3060 to 4060 range.
• A5xx: Mid-range (Arc A580). Budget 1080p gaming.
• A3xx: Entry-level (Arc A380). Light gaming only.

Intel’s Arc drivers have improved a lot since launch but still trail NVIDIA and AMD in older DirectX 9/10/11 games. If you see an Arc GPU in your system and old games run poorly, that’s a driver limitation, not a hardware defect.

What to Do After Identifying Your GPU

Knowing your GPU model unlocks four practical next steps:

1. Update Your Drivers

Outdated drivers cause crashes, poor FPS, and broken features. Go to nvidia.com/drivers, amd.com/en/support, or intel.com/download-center and enter your exact model. Game-ready drivers drop every few weeks and often include day-one optimizations for new releases.

2. Check Game Compatibility

Compare your GPU against a game’s system requirements on Steam or Epic. Sites like Can I Run It automate this comparison. If you’re between the minimum and recommended specs, expect playable FPS at medium settings.

3. Plan an Upgrade

GPU hierarchy charts from TechPowerUp and Tom’s Hardware rank every card by gaming performance. Find your current card, then look 2-3 tiers up for a meaningful jump. Moving from an RTX 3060 to an RTX 4070 gives roughly 70% more performance. Jumping just one tier (3060 to 3070) is often not worth the cost of upgrading.

4. Verify Used Card Purchases

If you bought a used GPU, run GPU-Z and compare the reported chip name against TechPowerUp’s database. Counterfeit cards are common on marketplace sites – a re-flashed GTX 960 will show “RTX 3060” in Device Manager but GPU-Z catches the mismatch immediately.

For maximum certainty, check the Hardware IDs. Open Device Manager, right-click your GPU, pick Properties, go to the Details tab, and set Property to “Hardware Ids”. You’ll see a string like PCIVEN_10DE&DEV_2684&SUBSYS_40811458&REV_A1. The VEN field is the vendor (10DE = NVIDIA, 1002 = AMD, 8086 = Intel). DEV identifies the chip. SUBSYS identifies the board partner (MSI, ASUS, Gigabyte). Look the full string up at pci-ids.ucw.cz – if the hardware ID doesn’t match the advertised card, walk away.

On Mac?

This guide focuses on Windows. On macOS, click the Apple menu in the top-left, pick About This Mac, then More Info, then System Report, then Graphics/Displays. You’ll see the chipset, VRAM, Metal support level, and connected displays. Apple Silicon Macs (M1/M2/M3/M4) report the unified-memory GPU that shares RAM with the CPU – there is no dedicated VRAM number.

Common Mistakes

Edge Cases and Platform Notes

Most of the methods above assume a standard Windows 10 or 11 install on x64 hardware with a GPU that’s physically sitting in the PCIe slot. Three scenarios don’t fit that mold, and they trip people up often enough to deserve their own notes.

Windows on ARM (Snapdragon X Elite)

Arm64 Windows 11 devices have Adreno GPUs that report through the Qualcomm driver stack. Task Manager and Settings both work and show the Adreno model name directly. msinfo32 may label the driver package differently from x64 systems. Avoid x86-only tools like older GPU-Z builds – use the Arm64-native version from TechPowerUp. Emulated x64 games don’t always pass GPU feature-level info correctly, so dxdiag may under-report the DirectX version the hardware actually supports. For the most reliable read on an ARM-based PC, use the vendor Qualcomm driver panel or the Settings app rather than relying on legacy tools.

WSL2 and virtual machines

In Windows Subsystem for Linux 2 with GPU pass-through enabled (wsl --update on recent builds), run nvidia-smi or rocm-smi inside the WSL shell to see the host GPU’s live utilization, VRAM use, and driver version. In Hyper-V or VMware guests you typically see a synthetic adapter like “Microsoft Hyper-V Video” or “VMware SVGA II” instead of the host GPU – these are virtual display adapters, not the real card the host is running. The same applies to RDP sessions: “Microsoft Remote Display Adapter” means you’re viewing a remote desktop stream, not the host GPU. Use the host machine directly to identify the physical card.

External GPUs (eGPU) and detection mismatches

Thunderbolt-connected eGPUs show in Device Manager under Display adapters when the enclosure is connected AND the vendor driver is installed. If the enclosure’s LED is on but no GPU appears in Device Manager, check three things: (1) Thunderbolt Control Center approved the device, (2) your BIOS/UEFI has Thunderbolt Security Level set to “User Authorization” or lower, and (3) the eGPU’s power supply is providing enough wattage for the card. If the card shows during BIOS POST but not in Windows, reinstall the vendor driver and reboot with the enclosure already connected – many eGPU enclosures behave badly with hot-plug after boot. If you’re troubleshooting a card that’s installed but not showing up at all, our Fix GPU Not Detected guide walks through the deeper diagnostics.

Which Method Should You Use?

Stop overthinking this. Match your situation to the right tool:

For 9 out of 10 people reading this, Task Manager is the only tool you need. Open it, click Performance, click GPU. You have your answer in under 15 seconds.

2026 update: The RTX 5070 and RX 9070 XT redefined mid-range GPU value. Our GPU guide covers current recommendations.

Frequently Asked Questions

How do I check what GPU I have on Windows 11?

Press Ctrl + Shift + Esc to open Task Manager, click Performance, then click GPU in the sidebar. Your graphics card name appears in the top-right. You can also open Settings, go to System, Display, Advanced display to see your GPU model. Both methods take under 15 seconds and need no downloads.

How do I check my GPU VRAM on Windows?

Open Task Manager (Ctrl + Shift + Esc), go to Performance, click GPU, and look for “Dedicated GPU Memory” at the bottom. That number is your actual VRAM. Don’t trust dxdiag’s “Approx. Total Memory” because it inflates the figure by adding shared system RAM. GPU-Z gives the most accurate reading including memory type.

Why does Windows show two GPUs?

Most PCs with a dedicated graphics card also have an integrated GPU built into the CPU. Task Manager labels them GPU 0 and GPU 1. The integrated chip handles desktop apps and web browsing to save power. The dedicated card activates for games and heavy workloads. This is normal behavior, not a hardware problem.

How do I check my GPU using Command Prompt or PowerShell?

Open PowerShell and run Get-CimInstance Win32_VideoController | Select Name. This prints every GPU in your system. In Command Prompt, use wmic path win32_VideoController get name. Add AdapterRAM, DriverVersion to the select list for more detail, but note that AdapterRAM caps at 4GB due to a WMI limitation.

How do I tell if my laptop GPU is integrated or dedicated?

Open Task Manager (Ctrl + Shift + Esc), click Performance. If you see two GPU entries, GPU 0 is usually the integrated chip (Intel UHD or AMD Radeon integrated) and GPU 1 is the dedicated card. If only one GPU appears and it shows an Intel UHD or generic AMD Radeon name, your laptop doesn’t have a dedicated GPU.

What does Microsoft Basic Display Adapter mean?

Windows can’t find proper drivers for your GPU and fell back to a generic software renderer. Games won’t run, and your display resolution is probably stuck at a low setting. Download the correct driver from NVIDIA, AMD, or Intel’s website and install it manually. If you don’t know which brand, run dxdiag before the drivers broke, or check the sticker on your laptop.

Can I check my GPU specs without installing anything?

Yes. Task Manager, dxdiag, Device Manager, Windows Settings, and PowerShell are all built into Windows 10 and 11. No downloads needed. Task Manager shows GPU name and VRAM. dxdiag shows driver version and export options. Only GPU-Z requires a download, and it’s optional – use it only if you need chip-level details or suspect a fake GPU.

How do I check my GPU using msinfo32?

Press Windows + R, type msinfo32, press Enter. Expand Components in the left pane, click Display. Your GPU shows up as “Adapter Description” with driver version, INF file, and installed drivers listed below. The Adapter RAM field caps at around 4 GB on cards with more VRAM – use Task Manager’s Dedicated GPU Memory for the real number.

Can two laptops with the same GPU name perform differently?

Yes, often by 20-30%. Laptop GPUs (like “RTX 4060 Laptop”) run at a Total Graphics Power the laptop vendor chooses, anywhere from 35W to 140W. A 140W RTX 4060 in a thick chassis beats an 80W RTX 4060 in a thin-and-light by a wide margin. GPU-Z’s “Default Board Power” field shows the configured TGP.

How do I identify an external GPU (eGPU) in Windows?

Plug in the Thunderbolt enclosure, approve it in Thunderbolt Control Center, then open Device Manager – the eGPU appears under Display adapters alongside your internal GPU. Task Manager Performance tab gives it its own pane. If the enclosure is powered but nothing shows, check BIOS Thunderbolt security level and reinstall the GPU vendor driver with the enclosure already connected.

Why does my graphics card show in BIOS but not in Windows?

Usually drivers are missing or corrupted. Device Manager will show “Microsoft Basic Display Adapter” instead of your real card. Fix: use Display Driver Uninstaller (DDU) in Safe Mode to scrub old drivers, then install the fresh driver from NVIDIA, AMD, or Intel’s website. If DDU clean-install doesn’t help, reseat the card in the PCIe slot and verify the 8-pin power connectors are fully seated.

How can I check my graphics card on Mac?

Click the Apple menu in the top-left, pick About This Mac, click More Info, then System Report, then Graphics/Displays. You’ll see the chipset, VRAM, Metal support, and connected displays. Apple Silicon Macs (M-series) report unified memory instead of dedicated VRAM because the GPU shares RAM with the CPU on one chip.

What’s the difference between a GPU and a graphics card?

A GPU is the processor chip – the silicon that does the math. A graphics card is the whole assembly – the GPU plus VRAM chips, cooler, circuit board, and HDMI/DisplayPort connectors. The terms are used interchangeably in casual talk. Integrated graphics (like Intel UHD or AMD Radeon Graphics) are GPUs without a separate card – the chip lives on the CPU.