https://amdai.net/updates Dated ROCm / llama.cpp / vLLM snapshots from the living tracker. https://amdai.net/updates#2026-10-12-consumer-stacks 2026-10-12-consumer-stacks Mon, 12 Oct 2026 00:00:00 GMT North star is now the two stacks from our test rig. llama.cpp Vulkan daily driver; Strata for Flash; /calculator and /stacks are first-class pages. https://amdai.net/updates#2026-10-12-rocm-10-1 2026-10-12-rocm-10-1 Mon, 12 Oct 2026 00:00:00 GMT 10.1.0 lists R9700 as gfx1201 and Radeon Ubuntu 26.04.1 / 24.04.5 / 22.04.5. Published ROCm CLI is v0.2.0, not a 1.0.0 tag. New pages for search, errors, backends, quants, RSS. https://amdai.net/updates#2026-09-22-vllm-0-30 2026-09-22-vllm-0-30 Tue, 22 Sep 2026 00:00:00 GMT Dated snapshot. AMD-validated vLLM 0.27 / ROCm 10.0.0 unchanged. Community image moved to vllm/vllm-openai-rocm:v0.30.0. Named Qwen defaults from that week were later withdrawn. https://amdai.net/updates#2026-09-21-weekly-refresh 2026-09-21-weekly-refresh Mon, 21 Sep 2026 00:00:00 GMT Dated snapshot. ROCm 10.0.0 unchanged for gfx1201. Reliability serve is AMD-aligned vLLM 0.27. vLLM 0.29 ROCm Docker remains the community cutting-edge pin (no newer stable that week). Named Qwen defaults from that week were later withdrawn. R9700 training KIs documented. https://amdai.net/updates#2026-09-09-vllm-0-29 2026-09-09-vllm-0-29 Wed, 09 Sep 2026 00:00:00 GMT Upstream vLLM 0.29 ships ROCm wheels and Docker. Extra index like wheels.vllm.ai/rocm/0.29.0/rocm723; image vllm/vllm-openai-rocm:v0.29.0. Plain pip install vllm is CUDA. AMD 10.0 docs still documented 0.27. https://amdai.net/updates#2026-09-02-rocm-7-14-1 2026-09-02-rocm-7-14-1 Wed, 02 Sep 2026 00:00:00 GMT 2 Sep 2026 patch on the 7.14 TheRock line. Still a valid pin if your wheel/image was built against it. Not a reason to install Ubuntu archive 7.1. https://amdai.net/updates#2026-08-26-rocm-10 2026-08-26-rocm-10 Wed, 26 Aug 2026 00:00:00 GMT Production TheRock line. R9700 / gfx1201 first-class. amdrocm-* packages, one gfx family, scoped /opt/rocm prefix. AMD’s notes pair PyTorch 2.13 and vLLM 0.27 — that is the supported combo, not a forever pin.