Local AI Registry

Recipe

qwen3-6-27b-bf16-rtx-pro-6000-blackwell-96gb-llama-cpp-tp1

qwen3-6-27b-bf16-rtx-pro-6000-blackwell-96gb-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
10355
Accelerators
1
Tensor parallel
1
Context tokens
131,072
Max concurrency
1
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/unsloth/Qwen3.6-27B-MTP-GGUF
Repository
unsloth/Qwen3.6-27B-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

Source
https://www.localmaxxing.com/en/runs/cmsovena500ddmp01rbj4jux0

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. /home/<USER>/models/quant/Qwen3.6-27B-Q6_K.gguf
  4. --alias
  5. Qwen3.6-27B-Q6_K
  6. --host
  7. 127.0.0.1
  8. --port
  9. 8095
  10. --ctx-size
  11. 131072
  12. --gpu-layers
  13. 99
  14. --flash-attn
  15. on
  16. --parallel
  17. 1
  18. --batch-size
  19. 2048
  20. --ubatch-size
  21. 1024
  22. --cache-type-k
  23. f16
  24. --cache-type-v
  25. f16
  26. --jinja
  27. --reasoning
  28. off
  29. --chat-template-kwargs
  30. {"enable_thinking":false}
  31. --temp
  32. 0.7
  33. --top-k
  34. 20
  35. --top-p
  36. 0.80
  37. --min-p
  38. 0.0
  39. --presence-penalty
  40. 1.5
  41. --repeat-penalty
  42. 1.0
  43. --spec-type
  44. draft-mtp
  45. --spec-draft-n-max
  46. 2
  47. --spec-draft-p-min
  48. 0.75
FlagValue
-m/home/<USER>/models/quant/Qwen3.6-27B-Q6_K.gguf
--aliasQwen3.6-27B-Q6_K
--host127.0.0.1
--port8095
--ctx-size131072
--gpu-layers99
--flash-attnon
--parallel1
--batch-size2048
--ubatch-size1024
--cache-type-kf16
--cache-type-vf16
--reasoningoff
--chat-template-kwargs{"enable_thinking":false}
--temp0.7
--top-k20
--top-p0.80
--min-p0.0
--presence-penalty1.5
--repeat-penalty1.0
--spec-typedraft-mtp
--spec-draft-n-max2
--spec-draft-p-min0.75

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
1131,0721,987.140.3282.3observedqwen3-6-27b-bf16-rtx-pro-6000-blackwell-96gb-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
rtx-pro-6000-blackwell-96gb
id
qwen3-6-27b-bf16-rtx-pro-6000-blackwell-96gb-llama-cpp-tp1
model instance id
unsloth-qwen3-6-27b-mtp-gguf--bf16
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-27b-bf16-rtx-pro-6000-blackwell-96gb-llama-cpp-tp1-sweep
status
candidate

capabilities

draft launch

accelerator backend
nvidia
arguments
-hf, unsloth/Qwen3.6-27B-MTP-GGUF:BF16, --n-gpu-layers, 999, --host, 0.0.0.0, --port, 8080, -c, 131072
container port
8,080
environment · LLAMA CACHE
/root/.cache/huggingface
host port
8,080
image
ghcr.io/ggml-org/llama.cpp:server-cuda12-b10481@sha256:b2497f8834f5ecb4e38530f6bf2734b8e0be107ff48e4720145911c86930f2ce
ipc
host
kind
docker
shm size
16g
synthesized · generated at
2026-08-31T22:12:17Z
synthesized · image provenance
gemma-4-12b-q4-k-m-rtx-3060-12gb-llama-cpp-tp1
synthesized · template
llama-cpp-server-v1

mounts

read onlytarget
No/root/.cache/huggingface

engine

name
llama.cpp
version
10355

serving

max concurrency
1
max context tokens
131,072
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
draft launch.entrypoint · provenance · captured at
2026-09-01T01:41:26Z
draft launch.entrypoint · reason linux-amd64-container-config-entrypoint
draft launch.entrypoint · state known
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
2026-09-01T10:20:39Z
serving.max concurrency · reason explicit-source-server-capacity
serving.max concurrency · state known
serving.max context tokens · provenance · captured at
2026-09-01T10:20:39Z
serving.max context tokens · reason explicit-source-context-limit
serving.max context tokens · state known

metadata

localmaxxing · backend
cuda
localmaxxing · batch size
1
localmaxxing · hardware label
RTX PRO 6000 Blackwell
localmaxxing · notes
Qwen3.6-27B BF16 GGUF benchmark with llama.cpp and MTP on NVIDIA RTX PRO 6000 Blackwell 96 GB. Single-user text completion, 131072 context, 128-token output cap, temperature 0, full GPU offload with MTP draft decoding. Recorded decode throughput 40.34 tok/s for 128 output tokens; 53 of 64 MTP draft tokens accepted; prompt tokens 561.
localmaxxing · observed command
llama-server -m /home/<USER>/models/quant/Qwen3.6-27B-Q6_K.gguf --alias Qwen3.6-27B-Q6_K --host 127.0.0.1 --port 8095 --ctx-size 131072 --gpu-layers 99 --flash-attn on --parallel 1 --batch-size 2048 --ubatch-size 1024 --cache-type-k f16 --cache-type-v f16 --jinja --reasoning off --chat-template-kwargs '{"enable_thinking":false}' --temp 0.7 --top-k 20 --top-p 0.80 --min-p 0.0 --presence-penalty 1.5 --repeat-penalty 1.0 --spec-type draft-mtp --spec-draft-n-max 2 --spec-draft-p-min 0.75
localmaxxing · run id
cmsovena500ddmp01rbj4jux0
localmaxxing · tokenized · arguments
llama-server, -m, /home/<USER>/models/quant/Qwen3.6-27B-Q6_K.gguf, --alias, Qwen3.6-27B-Q6_K, --host, 127.0.0.1, --port, 8095, --ctx-size, 131072, --gpu-layers, 99, --flash-attn, on, --parallel, 1, --batch-size, 2048, --ubatch-size, 1024, --cache-type-k, f16, --cache-type-v, f16, --jinja, --reasoning, off, --chat-template-kwargs, {"enable_thinking":false}, --temp, 0.7, --top-k, 20, --top-p, 0.80, --min-p, 0.0, --presence-penalty, 1.5, --repeat-penalty, 1.0, --spec-type, draft-mtp, --spec-draft-n-max, 2, --spec-draft-p-min, 0.75
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/cmsovena500ddmp01rbj4jux0 ↗