Version pins on this site are dated snapshots. Thehome status card is the only “what we would try this week.” If a note disagrees with the YAML behind that card, the YAML wins.
Release cadence
AMD publishes a release history tablewith version and date.[1]Recent production lines from that table:
| Version | Date on AMD’s table |
|---|---|
| 10.1.0 | 5 October 2026 |
| 10.0.0 | 26 August 2026 |
| 7.14.1 | 2 September 2026 |
| 7.14.0 | 15 July 2026 |
| 7.2.4 | 29 May 2026 |
The same table continues back through the 6.x and 5.x lines to 5.0.0 on 9 February 2022.[1]We do not invent a “every N weeks” promise from that list. Spacing is uneven (7.0.0 and 7.0.1 are consecutive days; 6.1.5 is dated after several 6.2.x and 6.3.x rows). Read the table, not a slogan.
Since ROCm 7.14, AMD uses TheRock as the build and release system.[2][5]That is why 10.x packages and prefixes do not look like the old apt install rocmarchive story. See the install tutorial.
Official vs “works”
- Official — the GPU, OS, or framework version appears on AMD’s10.1.0 compatibility matrix, the matching release notes, or the Linux system-requirementspage.[6][2][7]It is not a tok/s promise.
- Works — a community or first-party run that is not on that table, or that uses a newer upstream (for example a vLLM tag past AMD’s AI-ecosystem pin). We label those as tries. The version matrix says no datawhen we have not measured.
Nightly TheRock builds may list GPUs the production notes do not. The 10.1.0 notes say a GPU that is not listed might be community-enabled through TheRock nightlies.[2]That is still not “official.”
Distro support
Copy the Radeon row, not the Instinct row. From the 10.1.0 notes, Radeon Ubuntu support is:
- Ubuntu 26.04.1 — kernel GA 7.0
- Ubuntu 24.04.5 — kernel HWE 7.0
- Ubuntu 22.04.5 — kernel HWE 6.8
Those footnotes are in the 10.1.0 OS tables.[2]Instinct lists different kernels for the same Ubuntu numbers (for example 24.04.5 GA 6.8). Do not mix them. Other distros appear on Instinct or virtualization rows — we do not promote them for a first home Radeon without a cite.
ROCm 10.1 and the R9700
From the 10.1.0 release notes (5 October 2026) and the 10.1 blog:
- The Radeon AI PRO R9700 (and R9700S / R9600D / R9600) is on the hardware table as
gfx1201, RDNA 4, with the RX 9070 XT / GRE / 9070.[2] - 10.1.0 adds Radeon AI PRO R9600. The R9700 was already on the 10.0 device table (R9600 was not in that 10.0 R9000 row the same way).[2][3]
- AI ecosystem pins in 10.1.0: PyTorch 2.14.0, vLLM0.29.0. The notes say this replaces previous PyTorch 2.11.0 and vLLM 0.27.0 support.[2]
- AI ecosystem pins in 10.0.0 notes: PyTorch 2.13.0, vLLM0.27.0.[3]Those two snapshots disagree on the “previous” PyTorch number (10.1 says it replaced 2.11.0; 10.0 listed 2.13.0). We report both dated pages and do not pick a silent winner.
- LLVM 24, WSL2 technical preview,
amd-smias successor torocm-smi(deprecation period complete, per the blog).[4][2]
Go deeperTheRock packaging you will type (10.0 transition notes)
Official 10.0 TheRock notes: packages named amdrocm-* so they do not collide with leftover rocm packages; install prefix/opt/rocm/core-10.0 (confirm with ls /opt/rocm); installone GPU family, not every SKU.[5]RDNA 4 / R9700-shaped suffix is gfx1201 (AMD’s 10.0 table also mentions a combined gfx1200-gfx1201 suffix — apt-cache search on the day you install).
10.1 package names may differ. Copy fromthe 10.1.0 install page, not from memory.[10]
The ROCm CLI
The 10.1 blog introduces ROCm CLI v0.2.0 as a single-binary tool for installing, configuring, and running local AI on AMD GPUs, on Linux, Windows, and WSL2.[4]As of 2026-10-12 this site has not seen a v1.0.0 tag — checkGitHub releases rather than assuming a 1.0.[9]
Official getting-started commands:[8]
rocm— launcher (setup, serve, diagnose, chat, dashboard) on a TTY; status summary if piped.rocm install sdk— TheRock ROCm wheels and a matching PyTorch stack in a managed env.rocm serve qwen— built-in alias for a small assistant model. Not this site’s 35B daily driver.rocm examine/rocm diagnose/rocm fix— inspect and apply reported fixes.rocm version— CLI version plus the SDK and driver in use.
rocm install sdk next to an existing install shows up aslegacy_rocm_status: detected_unmanaged. The managed env is separate.[8]Do not assume the CLI replaced amd-smi or your llama.cpp build.
Version notes on this site
Historical snapshots — including ROCm 10.0 and a given vLLM tag asexamples of a week, not permanent law.
Oct 12, 2026 · labeled current that week
Consumer stacks — Qwen3.6 daily driver, Flash agent, calculator
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.
The site is a consumer AMD ROCm notebook, not a withdrawn-small-model tracker.
Daily driver on our test rig: Qwen3.6-35B-A3B Unsloth UD-Q5_K_XL, llama.cpp Vulkan, one Radeon AI PRO R9700, 262144 context, f16 KV, 0% system RAM offload.
Fast agent: Qwen 3.8 Flash Unsloth UD-Q6_K_XL on Strata v0.1.42 at 262K. Measured on that rig ~32.6 tok/s versus llama.cpp ~16 tok/s. llama.cpp Q6 is the fallback.
New pages: /stacks and the flagship /calculator. Exact serve flags and measured VRAM stay TODO until the rig owner fills them.
Oct 12, 2026 · labeled current that week
ROCm 10.1.0 notes, CLI v0.2.0, error paste, version matrix
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.
Cited from the 10.1.0 release notes and the 10.1 blog.
Install walkthrough: Install ROCm. Matrix cells that we have not measured say no data: /updates/matrix. RSS: /updates/rss.xml.
Sep 22, 2026 · archived snapshot
vLLM ROCm Docker try-tag v0.30.0
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.
Dated digest (2026-09-22, America/Chicago). ROCm 10.0.0 is still the R9700 (gfx1201) inference pin — release notes and the live compatibility matrix. AMD’s 10.0 ecosystem still lists vLLM 0.27.
Community cutting-edge Docker tag moved: vllm/vllm-openai-rocm:v0.30.0 (Docker Hub, pushed ~2026-09-22Z). Keep v0.29.0 as the previous try-tag (wheels.vllm.ai/rocm/0.29.0/rocm723 still documents that week). Do not mix 0.27 and 0.30 in one venv. Default PyPI vllm is still CUDA.
Named Qwen defaults from this week were later withdrawn. Qwen3.5 / 3.6 27B and 35B-A3B stay multi-GPU territory — a public llama.cpp multi-GPU writeup is not a single-card recipe.
Honesty unchanged: TORCH_BLAS_PREFER_HIPBLASLT=0 on training resets; SGLang AITER may fail; prefer llama.cpp HIP/Vulkan over Ollama on gfx1201.
Sep 21, 2026 · archived snapshot
Weekly refresh — vLLM 0.27 reliability (dated)
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.
This is a dated weekly card, not a new Core SDK line. ROCm 10.0.0 is still the R9700 inference pin (release notes). AMD’s 10.0 matrix still lists vLLM 0.27. Upstream vLLM 0.29.0 ROCm extras (vllm/vllm-openai-rocm:v0.29.0, wheels.vllm.ai/rocm/0.29.0/rocm723) stay as the community serve try — no newer stable ROCm vLLM found this morning.
Named Qwen defaults from this week were later withdrawn — do not treat them as current. Training/fine-tune on this SKU can GPU-reset (TORCH_BLAS_PREFER_HIPBLASLT=0); SGLang AITER may fail. Community training writeups often use ROCm 7.2.1 — keep inference on 10.0. TheRock gfx1201 perf logs are collection, not a bench table.
Sep 9, 2026 · archived snapshot
vLLM 0.29.0 ROCm extras
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.
This is the upstream snapshot people were running the week of 8–9 Sep 2026 — not AMD’s 10.0 “supported” 0.27 pair. Install with the ROCm extra index or the -rocm Docker tag. 0.30rc existed as bleeding. Re-check vLLM’s GPU install before you copy a wheel URL from memory.
Sep 2, 2026 · archived snapshot
ROCm 7.14.1 patch line
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.
7.14 is when TheRock went production. 7.14.1 is the 2 Sep 2026 patch train. If a container tag says rocm7.14.1, stay there on purpose. Do not “upgrade” by sprinkling 10.0 libraries into that image.
Aug 26, 2026 · archived snapshot
ROCm Core SDK 10.0.0
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.
AMD’s 10.0.0 notes (26 Aug 2026) put Radeon AI PRO R9700 on the device table as gfx1201. Framework table in that release: vLLM 0.27, SGLang 0.5.15, PyTorch 2.13. Use the versioned release notes and transition guide if you are reconstructing this week.
Living matrix (cited or “no data”): /updates/matrix. RSS:/updates/rss.xml.
Sources
Every version and support claim on this page points here. If a sentence has no mark, treat it as lab opinion or a TODO — not as AMD policy.