Recurring crashes since initial release.

Hello, I have had significant stability issues since launch back in 2025, but I did NOT have these problems in the open beta. I stopped playing in November 2025 because of it, but am trying to get back into the game. I have been troubleshooting the crashes with my good friend (Claude) and asked them to summarize everything I've tried so far. Here is Claude's summary:

Symptoms:
Since launch in October 2025, the game crashes to the desktop at random, anywhere from about 10 seconds to an hour after starting. It happens in menus and in matches. I also get freezes where the world simulation stops but I can still move the camera and the foliage keeps animating. The game never recovers.

What I’ve ruled out:

• RAM: MemTest86, 4 full passes at 3990 MT/s, 0 errors. Crashes continue at JEDEC-level speeds.
• CPU: Intel Processor Diagnostic Tool passed. OCCT CPU test, no errors. No WHEA errors in Event Viewer. Latest BIOS with current microcode. Crashes continue with turbo disabled.
• Storage: Moved the game to a known healthy NVMe drive. No disk errors in the System log.
• Software: Steam and NVIDIA overlays off, Razer software closed, clean GPU driver install, game files verified.

Crash dump analysis (38 dumps, Oct 2025 to Oct 2026):
Every crash is an EXCEPTION_ACCESS_VIOLATION inside bf6.exe itself, except one E_OUTOFMEMORY (0x8007000E) raised from the DX12 render code. The crashes keep landing on the same few code sites, which shift slightly with each game build. That points to a software bug rather than random hardware corruption.

• Stall dump captured (build 6abd4eb0, 10/11 10:34): The Heartbeat thread force-crashed the process (0xDEAFBABE write at bf6.exe+0x271c059), followed by the 0xBAD57411 termination at +0x11b0737. At the moment of the stall, GameSimulationLoop and MainLoop were blocked waiting on the job system, 26 of 28 Job threads were idle-waiting, and only Job0 (in a hash-table lookup at +0x5a4c2de) and Job5 (in the scheduler at +0x403dca2) were running. The DirectStorage, D3D, network and audio threads were all idle. This looks like a job-system dependency or wakeup bug rather than an I/O or GPU stall. GameStall dumps are usually written as 0 KB; this one is complete and attached.
• Read through an invalid -1 pointer (17 crashes). Offsets shown with the bf6.exe PE timestamp:
  • build 68dbf456: +0x582c4f3, +0x582caff
  • build 68ecb235: +0x582e8df (×3), +0x582e2d3 (×3)
  • build 68f1dd5f: +0x5890193 (×8)
  • build 69024468: +0x589a843
• Virtual call through a freed object while iterating a callback/listener list (5 crashes). The code walks a pointer array, calls [vtable+8] on each entry, then does a lock xadd decrement. The object’s vtable pointer is garbage, which looks like use-after-free or a race.
  • build 68dbf456: +0x548dfc2 (×3)
  • build 68ecb235: +0x548fda2
  • build 6abd4eb0: +0x5bbcd02, on worker thread “Job3”
• Write to a null pointer (4 crashes).
  • build 68dbf456: +0x56f2b45 (×2)
  • build 68ecb235: +0x56f4925, +0x56f4958
• Recent one-offs, build 6abd4eb0:
  • +0x642c0ed (thread “Job0”): read past the end of an allocation
  • +0x508e4c0 (thread “Job3”): read from a freed object, render code on the stack

The listener-list crash is identical machine code in October 2025 and October 2026, so this bug has survived every patch since launch. The crashing threads are job-system workers. Combined with the random timing, this looks like a threading race, which a 32-thread CPU may hit more often.

I've attached all of the dumps as a .zip. I hope this helps to resolve this issue for me and anyone else experiencing it!