CPU Temp Reading 120°C at Idle? Here's the Real Fix

Hardware – RAM & MB Intermediate 👁 10 views 📅 Jun 15, 2026

Your CPU hitting 120°C at idle? Usually a software bug, sensor glitch, or misreading. Here's how to tell if it's real or fake.

You open HWMonitor or your BIOS and see your CPU is at 115°C. While the PC is doing nothing. Maybe you're on a Ryzen 7 3700X with a stock cooler, or an Intel i5-12400 with a cheap air tower. You touch the heatsink and it's barely warm. Something's wrong.

I've seen this exact thing maybe a dozen times in the last couple years. Last month a small law office called because their Dell Optiplex kept shutting down. Customer swore the CPU was „melting.“ I touched the case – ice cold. Turned out a BIOS update had borked the sensor readout. Another time, a client's Ryzen 9 showed 127°C in Ryzen Master but ran perfectly fine. So let's cut the panic and figure out what's actually happening.

Why This Happens (Two Main Culprits)

You're almost certainly dealing with one of these:

  • Software misreading the sensor – The most common cause. Old versions of CoreTemp, HWMonitor, or Speccy read the wrong sensor. On newer AMD Ryzens (5000 and 7000 series), the CPU has multiple thermal sensors – one on the die, one near the I/O die, and sometimes a virtual sensor that UEFI uses. Software can grab the wrong one, especially if it's outdated.
  • Bad sensor or motherboard glitch – Less common but real. A bent pin on the CPU socket, a bit of thermal paste in the wrong spot, or a flaky voltage regulator can send false readings. On Intel 12th/13th gen, the i9s run hot, but even they don't hit 115°C at idle unless something's badly wrong.
  • BIOS bug after update – Seen this on ASUS and MSI boards especially. A BIOS flash can corrupt the thermal table the motherboard uses to report temps.

The Fix: Step-by-Step

Skip the panic. Don't reapply thermal paste yet. Do this first:

  1. Check with a second tool. Download HWiNFO64 (free, no bloat). Run it in „Sensors Only“ mode. Look for CPU (Tctl/Tdie) or CPU Package. If HWiNFO shows 40°C while HWMonitor shows 120°C, it's a software bug. Uninstall the old tool and stick with HWiNFO.
  2. Check in BIOS. Reboot and go straight into BIOS (usually Delete or F2). Look for CPU temperature or PC Health. BIOS reads directly from the chip – no software layer. If BIOS shows 35-45°C at idle, your hardware is fine. The issue is Windows-side.
  3. Update your monitoring software. If you're using CoreTemp 1.17 from 2019, that's your problem. AMD and Intel both changed sensor behavior over the years. Grab the latest version of HWiNFO or Ryzen Master / Intel XTU.
  4. Reset BIOS to defaults. If BIOS itself shows a crazy temp (like 110°C), load optimized defaults. On ASUS boards, I've seen a custom fan curve or voltage offset cause the sensor to report wrong. Load defaults, save, reboot.
  5. Re-seat the CPU cooler. If you've done all the above and HWiNFO still shows 90°C+ at idle, then it's time to get physical. Shut down, unplug, take the cooler off. Clean old thermal paste with isopropyl alcohol. Reapply a pea-size dot of Arctic MX-6 or Noctua NT-H2. Re-mount the cooler evenly – tighten in a cross pattern, not too hard (don't crack the CPU).

What If It's Still Wrong?

If after all that you're still seeing 100°C+ at idle, check these:

  • Check for a bent pin or socket damage. On AMD AM4/AM5, a bent pin can cause a false temp reading. On Intel LGA 1700, look for bent socket pins. I had a client whose kid dropped a screw on the socket – one pin was bent, temp sensor went haywire.
  • Check if the actual CPU cooler is making contact. Some coolers, especially cheap ones, have mounting brackets that are easy to install backwards. The pump block won't sit flat. You'll feel uneven heat on the heatpipes.
  • Check power supply voltage. A dying PSU can send unstable voltage to the VRMs, which can confuse the thermal sensor. Unlikely but possible.
  • Board replacement time. If the sensor on the CPU die itself is dead (rare but I've seen it on a Ryzen 3 3200G), the only fix is a new CPU. But test the board with a different CPU first if you have one.

Bottom line: 99% of the time it's a software misread. Don't go buying thermal paste or a new cooler until you've checked with HWiNFO and your BIOS. If the heatsink is cool to the touch and the PC isn't throttling, you're fine. If it's actually hot and throttling, then we have a real problem.

Was this solution helpful?