From 4412626681fa655f16f8846034ff3a0810b3f67b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?J=C3=BCrgen=20Schmied?=
 <600316+jschmied@users.noreply.github.com>
Date: Tue, 20 Oct 2026 23:23:33 +0200
Subject: [PATCH] docs: the 27B recipe names its native CUDA backend and which
 formats it reads

The page said the 27B has no CUDA backend; 1.1.0's native engine serves the MLX affine checkpoints
on CUDA. Its CUDA section, which promises NVFP4 and EXL3 checkpoints, describes the Python engine
(0.6.6); it now says so, and that neither reader is ported to the native engine yet.
---
 docs/recipes/qwen3.8-27b.md | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/docs/recipes/qwen3.8-27b.md b/docs/recipes/qwen3.8-27b.md
index a4b93279af..acd1afa3d3 100644
--- a/docs/recipes/qwen3.8-27b.md
+++ b/docs/recipes/qwen3.8-27b.md
@@ -2,7 +2,9 @@
 
 The native engine serves `TensorFold/Qwen3.8-27B-MLX-4bit` on Metal, qualified on M5 Max and M3 Ultra, from
 `zig/src/families/qwen27/`. It drafts with DFlash2 (`zig/src/families/qwen27/dflash/`) only when `--drafter` names
-the drafter; without it the model decodes plain. It has no CUDA backend yet.
+the drafter; without it the model decodes plain. On CUDA the native engine serves the MLX affine checkpoints (2, 3,
+4, 5, 6 and 8 bits in groups of 64) from `zig/src/families/qwen3_5/cuda_*.zig`, with DFlash2 drafts and
+`--kv-cache fp8`; it reads no NVFP4 or EXL3 checkpoint yet.
 
 ```bash
 tensorfold pull TensorFold/Qwen3.8-27B-MLX-4bit
@@ -83,6 +85,10 @@ quality comparisons are TBD [release-0.3.5].
 
 ## CUDA
 
+This section is the Python engine's (0.6.6). The native CUDA engine reads only MLX affine checkpoints so far: the
+NVFP4 and EXL3 readers below have not been ported to it (1.1.0 carries the shared EXL3 module of #832, which the
+27B does not read through yet).
+
 On CUDA the 27B serves NVFP4 and EXL3 checkpoints, and the MLX 4-bit checkpoint as the portable option: the same
 files a Mac serves, and the only format two ranks read. `tensorfold serve` loads the checkpoint you name; it picks
 none by itself. Use the [CUDA container setup](https://github.com/ashhart/TensorFold/blob/v0.6.6/RUNBOOK.md#nvidia-gpus)