Local AI Registry

Recipe

qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1

qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
JSON APIFind compatible recipes

Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.

Record

Status
candidate
Source
localmaxxing
Engine
llama.cpp
Engine version
llama.cpp
Accelerators
1
Tensor parallel
1
Context tokens
500,000
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/GestaltLabs/Qwen3.6-35B-A3B-NSC-ACE-SABER-GGUF-MTP
Repository
GestaltLabs/Qwen3.6-35B-A3B-NSC-ACE-SABER-GGUF-MTP
Status
known
Link type
Exact Hub repository

Public Hugging Face repository confirmed by the Hub API.

Candidate: useful compatibility or speed evidence. The registry does not offer Run until promotion requirements are met.

Observed configuration

llama.cpp

Evidence only · candidate · reference

Candidate evidence — not a Run contract

Source
https://www.localmaxxing.com/en/runs/cmq45u04k003lmu01cnd8jgzt

Observed source tokens

Mechanical split of the source command. Unverified against the engine CLI. The registry does not offer Run for this recipe.

  1. llama-server
  2. -m
  3. Qwen3.6-35B-A3B-NSC-ACE-SABER-ROCmFP4-STRIX_LEAN.gguf
  4. -c
  5. 500000
  6. -b
  7. 32768
  8. -ub
  9. 2048
  10. -t
  11. 32
  12. --threads-batch
  13. 16
  14. --cache-ram
  15. 4096
  16. --ctx-checkpoints
  17. 0
  18. --cache-type-k
  19. f16
  20. --cache-type-v
  21. f16
  22. --spec-type
  23. draft-mtp
  24. --spec-type
  25. ngram-mod
  26. --draft-max
  27. 6
  28. --no-mmap
  29. --kv-unified
  30. --no-cont-batching
  31. --no-cache-idle
FlagValue
-mQwen3.6-35B-A3B-NSC-ACE-SABER-ROCmFP4-STRIX_LEAN.gguf
-c500000
-b32768
-ub2048
-t32
--threads-batch16
--cache-ram4096
--ctx-checkpoints0
--cache-type-kf16
--cache-type-vf16
--spec-typedraft-mtp
--spec-typengram-mod
--draft-max6

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
1500,000772137.9—observedqwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

hardware count
1
hardware id
ryzen-ai-max-plus-395-128gb
id
qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
model instance id
gestaltlabs-qwen3-6-35b-a3b-nsc-ace-saber-gguf-mtp--rocmfp4
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep
status
candidate

capabilities

engine

name
llama.cpp
version
llama.cpp

serving

max context tokens
500,000
tensor parallel
1
Provenance & metadata (3)

facts

capabilities.chat · provenance · captured at
2026-08-30T09:10:02Z
capabilities.chat · reason capability-not-verified
capabilities.chat · state unknown
capabilities.reasoning · provenance · captured at
2026-08-30T09:10:02Z
capabilities.reasoning · reason capability-not-verified
capabilities.reasoning · state unknown
capabilities.tools · provenance · captured at
2026-08-30T09:10:02Z
capabilities.tools · reason capability-not-verified
capabilities.tools · state unknown
capabilities.vision · provenance · captured at
2026-08-30T09:10:02Z
capabilities.vision · reason capability-not-verified
capabilities.vision · state unknown
engine.graph mode · provenance · captured at
2026-08-30T09:10:02Z
engine.graph mode · reason runtime-detail-not-published
engine.graph mode · state unknown
serving.kv cache tokens · provenance · captured at
2026-08-30T09:10:02Z
serving.kv cache tokens · reason kv-cache-capacity-not-published
serving.kv cache tokens · state unknown
serving.max concurrency · provenance · captured at
2026-08-31T23:03:15Z
serving.max concurrency · reason server-capacity-not-evidenced
serving.max concurrency · state unknown

metadata

localmaxxing · backend
rocm
localmaxxing · batch size
32,768
localmaxxing · hardware label
Ryzen AI Max 395
localmaxxing · notes
Evidence: eval=137.91 tok/s; output=50 tokens.
localmaxxing · observed command
llama-server -m Qwen3.6-35B-A3B-NSC-ACE-SABER-ROCmFP4-STRIX_LEAN.gguf -c 500000 -b 32768 -ub 2048 -t 32 --threads-batch 16 --cache-ram 4096 --ctx-checkpoints 0 --cache-type-k f16 --cache-type-v f16 --spec-type draft-mtp --spec-type ngram-mod --draft-max 6 --no-mmap --kv-unified --no-cont-batching --no-cache-idle
localmaxxing · run id
cmq45u04k003lmu01cnd8jgzt
localmaxxing · tokenized · arguments
llama-server, -m, Qwen3.6-35B-A3B-NSC-ACE-SABER-ROCmFP4-STRIX_LEAN.gguf, -c, 500000, -b, 32768, -ub, 2048, -t, 32, --threads-batch, 16, --cache-ram, 4096, --ctx-checkpoints, 0, --cache-type-k, f16, --cache-type-v, f16, --spec-type, draft-mtp, --spec-type, ngram-mod, --draft-max, 6, --no-mmap, --kv-unified, --no-cont-batching, --no-cache-idle
localmaxxing · tokenized · fidelity
faithful

provenance

captured at
2026-08-30T09:10:02Z

sources

captured atkindurl
2026-08-30T09:10:02Znormalized-recipewww.localmaxxing.com/en/runs/cmq45u04k003lmu01cnd8jgzt ↗