On 2026-10-30 we asked 42 people who publish OpenCore EFIs for Intel 11th to 14th gen and Core Ultra machines for their iGPU details. By 05:00 EDT on 2026-10-31, 14 had answered: 12 on their own repos and 2 as probe reports in WTLIntelGfx #3. Everything below is what they reported. We have not reproduced any of it.
| Machine | iGPU | macOS | Result | Link |
|---|---|---|---|---|
| Lenovo Legion Y9000P 2022, i7-12700H | 8086:46A6 rev 0C | 26.7.1 (25G241) | stage 2 passes on 0.0.46, pipe A lit | #3 |
| HP Victus 15-fa0xxx, i7-12650H | 8086:46A3 rev 0C | 26.7.1 (25G241) | stages 1 to 3 pass on 0.0.38, stage 4 panics | #3 |
The stage 4 panic is a NULL format string in xg_vsnprintf, reached from kobject_init_and_add when intel_detect_pch asks LinuxKPI for the ISA bridge. The bridge's newbus device probably has no name yet, so the kobject gets a NULL name. Arrow Lake does not take this path, which is why our two machines never hit it. Confidence: Medium (backtrace and source read, not reproduced). Nothing between 0.0.39 and 0.0.46 touches it.
| Machine | iGPU | Class | BIOS mode | Notes | Link |
|---|---|---|---|---|---|
| Dell G15 5511, i7-11800H | 8086:9A60 rev 01 (Tiger Lake-H UHD) | 030000 in Hybrid | MUX | Discrete removes the iGPU from ACPI and PCI. Hybrid leaves it on IONDRVFramebuffer, and the RTX 3060 keeps Metal 3 as a compute-only GPU | link |
| Gigabyte Gaming A16, i7-13620H | 8086:A7A8 rev 04 | 038000 | Discrete, disable-gpu | link | |
| Lenovo Legion Y7000P IRX10, i7-14650HX | 8086:A788 (UHD, 32 EU) | 038000 | Discrete, nothing blocked | Tahoe 26.7.2 | link |
| ROG Strix G16 G614JVR, i9-14900HX | 8086:A788 | display | dGPU only / Dynamic | Tahoe 26.7.2 | link |
| Mechrevo GM6AR57C, i7-12800HX | 8086:4688 rev 0C (UHD, 32 EU) | 030000 | Hybrid | WhateverXe 0.6.1 on AppleIntelTGLGraphics, Metal 2, macOS 15.8.2 | link |
| Acer Nitro AN515-58, i5-12500H | 8086:46A6 rev 0C (80 EU) | 030000 | Optimus | From Linux; macOS does not boot there yet | link |
Two things came out of this table.
- In Discrete MUX mode the iGPU either disappears or shows up as class 0x0380 (display, other) instead of 0x0300. The XeGfx stage 2 probe only looks for subclass 00, so it would miss the A16 and the Y7000P as configured. The kexts match on base class 03 and are fine. The A16 owner found this by reading
xg_probe. Confidence: High (code read). - 8086:9A60 (Tiger Lake-H) and 8086:A788 (Raptor Lake-HX, the small 32 EU part) are not in the XeGfx README's ID table yet.
victorwitkamp/NootedGreen, a fork of sgiammori/NootedGreen with a raptor-lake-hx branch. Its 2026-10-30 commits port i915 pieces (forcewake, pipe VDSC, render engine workarounds, EC backlight). The owner says graphics work with Metal on his ROG Strix G814JI (comment). NootedGreen spoofs Tiger Lake for Apple's driver, which is a different route from ours. Confidence: Unverified. Ideas only, no code.
- Willing, nothing sent yet: 7HE-W0R1D (Z890M, Arrow Lake desktop, this weekend), nidemimihihi (i7-14650HX, wants step by step), cells114514 (info only, no test drivers on his main machine).
- Declined: two owners, one who did not want to share system details and one who no longer has the machine.