Crown Dynamic MTP v7 is a custom mixed-precision GGUF quantization of unsloth/Qwen3.6-35B-A3B-MTP-GGUF, tuned for local AMD Strix Halo serving with llama.cpp native MTP speculative decoding.
This is not a generic cloud inference recipe. It is a Strix Halo owner profile: single-user, high-context, Vulkan, native MTP, split mode none, flash attention, vision projector support, and polling settings chosen for this machine class.
mmproj-F16.mmproj is the Qwen3.6 35B A3B GGUF-format vision projector. It is stored with a .mmproj repo extension so Hugging Face's GGUF parser keeps the main card focused on the 35B language model rather than the smaller projector.
Technical Metadata
Hugging Face may round the parsed GGUF tensor count to 36B in its automatic badge. This release is the Qwen3.6 35B-A3B MoE family: about 35B-class total parameters with roughly 3B active parameters per token.
Field
Value
model size
35B-A3B MoE
total parameters
35B class
active parameters
~3B class
architecture
qwen35moe
GGUF size label
35B-A3B
direct upstream GGUF
unsloth/Qwen3.6-35B-A3B-MTP-GGUF
base family
Qwen/Qwen3.6-35B-A3B
local runtime format
mixed-precision Vulkan GGUF
vision projector
mmproj-F16.mmproj
What This Is
Crown v7 is not a renamed upstream quant. It uses a custom mixed-precision tensor recipe:
MXFP4 on blocks 0-27 routed expert gate/up tensors
Q4_K_M fallback on blocks 28-35 routed expert gate/up tensors
Q5_K on blocks 36-39 routed expert gate/up tensors
Q5_K / Q6_K for routed down experts
Q8_0 for attention, shared experts, token embeddings, output, and selected MTP tensors
The full recipe is included in recipes/halo-mtp-dyn-v7.md and recipes/halo-mtp-dyn-v7.tensor-types.txt.
Strix Halo Serving Profile
Reference profile:
AMD Ryzen AI MAX / Radeon 8060S Strix Halo class machine
The exact public Strix profile is included as model-profiles/qwen3.6-35b-a3b-crown-halo-mtp-dynamic.env, and the serving script is included as scripts/serve_halo_mtp_dyn_v7.sh.
For text-only use, you may omit --mmproj. For image input, keep mmproj-F16.mmproj next to the main GGUF and pass it with --mmproj.
KV Cache Choice
Use f16 target KV and f16 draft KV with b2048/u512.
Older local comparison rows included q8_0/q8_0 KV, but that is not the public Strix Halo recommendation for this release. The benchmark story and serving recipe are centered on the f16/f16 MTP profile.
Where It Shines On Strix Halo
The point of this release is native MTP behavior under compatible llama.cpp serving. It shines when the MTP head gets accepted at a high rate, especially structured or repetitive long-decode workloads.
On structured 4k-token prompt / 256-token generation tests:
setting
workload
generation tok/s
MTP acceptance
accepted / drafted
MTP depth 4, b2048/u512, f16 KV
JSON
105.79
97.6%
203 / 208
MTP depth 4, b2048/u512, f16 KV
code
106.13
99.5%
203 / 204
MTP depth 4, b2048/u512, f16 KV
chat
104.67
97.6%
203 / 208
MTP depth 4, b2048/u512, f16 KV
tool-call JSON
104.97
97.6%
203 / 208
MTP depth 2, b2048/u512, f16 KV
JSON
91.08
100.0%
170 / 170
MTP depth 2, b2048/u512, f16 KV
code
90.59
100.0%
170 / 170
MTP depth 2, b2048/u512, f16 KV
chat
90.07
100.0%
170 / 170
MTP depth 2, b2048/u512, f16 KV
tool-call JSON
90.60
100.0%
170 / 170
Best observed structured-decode slice: MTP depth 4, roughly 105-106 tok/s with 97-99% draft-token acceptance.
A separate context sweep found a strong mid-context decode row:
context
setting
prompt tokens
generated
generation tok/s
MTP acceptance
32k
MTP depth 4
20,263
128
79.58
86.0%
32k
MTP depth 2
20,263
128
59.74
66.1%
Vision Smoke Test
Vision was validated locally on June 15, 2026 with the normal Crown Dynamic profile, the mmproj-F16 projector, and native MTP enabled. Test image:
/srv/desktop-data/cirudata/ciruoutfit.png
Prompt:
Look at this image. In one short sentence, name the dominant hair color and the object covering the character's ear.
Response:
The dominant hair color is green, and the object covering the character's ear is a cybernetic headset.
The OpenAI-compatible llama.cpp response included MTP draft timing fields: draft_n=24 and draft_n_accepted=19.
Strix Owner Notes
--parallel 1 is intentional. MTP is the constraint, and this profile is for single-user local serving.
-c 131072 is intentional. This release is focused on high-context Strix Halo use.
Use --spec-type mtp; plain GGUF loading will not show the behavior this model was built for.
Use the polling settings from the profile: --poll 100, --poll-batch 1, --spec-draft-poll 1, and --spec-draft-poll-batch 1.
b2048/u512 is the recommended Strix MTP default from the acceptance matrix; larger b8192/u512 did not improve acceptance and added latency.
f16/f16 KV is the recommended MTP setting for Strix Halo.
Use --mmproj mmproj-F16.mmproj for image input. Omit it only for text-only serving.
Important Benchmark Caveat
This is not a universal "always faster" GGUF.
In ordinary non-speculative llama-bench, Crown v7 looks similar to other strong Qwen3.6 35B A3B quants. The point of this release is the MTP behavior under compatible Strix Halo llama.cpp serving, not raw no-speculative token generation.
Local context-sweep testing showed:
MTP helped most on structured/repetitive decode workloads with high draft acceptance.
MTP depth 4 was the strongest tested setting for structured decode.
Very long active contexts, especially around 64k, did not show an end-to-end MTP win on this Vulkan host.
If you run this as a normal GGUF without MTP support, expect a strong Q4-size mixed quant, not the accelerated MTP profile.
