Recipe
qwen3-6-35b-a3b-ud-q4-k-xl-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
qwen3-6-35b-a3b-ud-q4-k-xl-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.
Record
- Status
- candidate
- Source
- localmaxxing
- Engine
- llama.cpp
- Engine version
- toolbox-rocm-7.2.4-post-gtt
- Accelerators
- 1
- Tensor parallel
- 1
- Context tokens
- 8,192
- chat
- unknown
- reasoning
- unknown
- tools
- unknown
- vision
- unknown
Hugging Face model card
Identity
https://huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF- Repository
- unsloth/Qwen3.6-35B-A3B-MTP-GGUF
- 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
No tokenized launch fields. This is measured or documented compatibility, not a Docker launch.
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status | Sweep |
|---|---|---|---|---|---|---|
| 1 | 8,192 | 894.2 | 75.9 | 8,653 | observed | qwen3-6-35b-a3b-ud-q4-k-xl-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-ud-q4-k-xl-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
- model instance id
- unsloth-qwen3-6-35b-a3b-mtp-gguf--ud-q4-k-xl
- recipe source
- localmaxxing
- schema version
- local-ai-registry/v1
- speed sweep ids
- qwen3-6-35b-a3b-ud-q4-k-xl-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep
- status
- candidate
capabilities
engine
- name
- llama.cpp
- version
- toolbox-rocm-7.2.4-post-gtt
serving
- max context tokens
- 8,192
- 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
- 1
- localmaxxing · hardware label
- Ryzen AI Max 395
- localmaxxing · notes
- Post-GTT/UMA tuning Horizon benchmark. BIOS UMA=1GB; active kernel cmdline includes amd_iommu=off amdgpu.gttsize=126976 ttm.pages_limit=32505856. ROCm sees Radeon 8060S as 126976 MiB; sysfs GTT total 124 GiB. Measured on existing Docker container kyuz0/amd-strix-halo-toolboxes:rocm-7.2.4, endpoint 8002. Scenario: 3 streamed chat-completion runs, ctx=8192, target ~7.5k content tokens, max_tokens=128, temperature=0. Reported numbers are means across runs. Native prompt tokens per run: 7702, 7707, 7705. tokSPrefill is native llama.cpp metrics delta prompt_tokens_total / prompt_seconds_total; tokSOut is native tokens_predicted_total / tokens_predicted_seconds_total; TTFT is wall-clock first SSE chunk. Run-specific nonce at beginning of user content prevents prompt/KV cache reuse. Server flags: --ctx-size 8192 -ngl 999 -fa 1 --no-mmap --jinja --metrics --mmproj mmproj-F16.gguf --reasoning-format auto --spec-type draft-mtp --spec-draft-n-max 3 --spec-draft-p-min 0.75 --no-webui. Model file: Qwen3.6-35B-A3B-UD-Q4_K_XL.gguf. Local artifacts: /home/joshua/benchmarks/horizon-post-gtt-benchmark-2026-06-29/summary.json
- localmaxxing · run id
- cmqzjqwky02p7oe01jpx9fes5
provenance
- captured at
- 2026-08-30T09:10:02Z
sources
| captured at | kind | url |
|---|---|---|
| 2026-08-30T09:10:02Z | normalized-recipe | www.localmaxxing.com/en/runs/cmqzjqwky02p7oe01jpx9fes5 ↗ |