Local AI Registry

Recipe

ornstein3-6-27b-mtp-nsc-ace-saber-q4-k-m-rtx-3090-24gb-llama-cpp-tp1

ornstein3-6-27b-mtp-nsc-ace-saber-q4-k-m-rtx-3090-24gb-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
buun-llama-cpp (86-real, MSVC 19.31, CUDA 12.8)
Accelerators
1
Tensor parallel
1
Context tokens
262,144
Max concurrency
1
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/GestaltLabs/Ornstein3.6-27B-MTP-NSC-ACE-SABER-GGUF
Repository
GestaltLabs/Ornstein3.6-27B-MTP-NSC-ACE-SABER-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/cmpcxbhcw00fynw017wgjyvsp

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. --threads
  3. 12
  4. --prio
  5. 3
  6. --n-predict
  7. 32768
  8. --batch-size
  9. 256
  10. --ubatch-size
  11. 128
  12. --flash-attn
  13. on
  14. --cache-type-k
  15. turbo3_tcq
  16. --cache-type-v
  17. turbo2_tcq
  18. --no-mmap
  19. --n-gpu-layers
  20. all
  21. --fit
  22. off
  23. --ctx-size
  24. 262144
  25. --model
  26. Ornstein3.6-27B-MTP-NSC-ACE-SABER-Q4_K_M-MTP.gguf
  27. --cache-type-k-draft
  28. turbo3_tcq
  29. --cache-type-v-draft
  30. turbo2_tcq
  31. --temperature
  32. 0.6
  33. --top-p
  34. 0.95
  35. --min-p
  36. 0.01
  37. --repeat-penalty
  38. 1.0
  39. --presence-penalty
  40. 0.0
  41. --ctx-checkpoints
  42. 8
  43. --cache-ram
  44. 8192
  45. --kv-unified
  46. --parallel
  47. 1
  48. --mmproj
  49. mmproj-Ornstein3.6-27B-MTP-NSC-ACE-SABER-F16.gguf
  50. --no-mmproj-offload
  51. --image-min-tokens
  52. 1024
  53. --alias
  54. Qwen3.6-27B
  55. --props
  56. --reasoning
  57. on
  58. --chat-template-kwargs
  59. {"preserve_thinking":true}
  60. --reasoning-budget
  61. 8192
  62. --reasoning-budget-message
  63. . My reasoning budget is exhausted, but I have enough information to answer directly now.
  64. --spec-type
  65. draft-mtp
  66. --spec-draft-n-max
  67. 2
  68. --spec-draft-p-min
  69. 0.75
FlagValue
--threads12
--prio3
--n-predict32768
--batch-size256
--ubatch-size128
--flash-attnon
--cache-type-kturbo3_tcq
--cache-type-vturbo2_tcq
--n-gpu-layersall
--fitoff
--ctx-size262144
--modelOrnstein3.6-27B-MTP-NSC-ACE-SABER-Q4_K_M-MTP.gguf
--cache-type-k-draftturbo3_tcq
--cache-type-v-draftturbo2_tcq
--temperature0.6
--top-p0.95
--min-p0.01
--repeat-penalty1.0
--presence-penalty0.0
--ctx-checkpoints8
--cache-ram8192
--parallel1
--mmprojmmproj-Ornstein3.6-27B-MTP-NSC-ACE-SABER-F16.gguf
--image-min-tokens1024
--aliasQwen3.6-27B
--reasoningon
--chat-template-kwargs{"preserve_thinking":true}
--reasoning-budget8192
--reasoning-budget-message. My reasoning budget is exhausted, but I have enough information to answer directly now.
--spec-typedraft-mtp
--spec-draft-n-max2
--spec-draft-p-min0.75

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
1262,14498.932.9—observedornstein3-6-27b-mtp-nsc-ace-saber-q4-k-m-rtx-3090-24gb-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-3090-24gb
id
ornstein3-6-27b-mtp-nsc-ace-saber-q4-k-m-rtx-3090-24gb-llama-cpp-tp1
model instance id
gestaltlabs-ornstein3-6-27b-mtp-nsc-ace-saber-gguf--q4-k-m
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
ornstein3-6-27b-mtp-nsc-ace-saber-q4-k-m-rtx-3090-24gb-llama-cpp-tp1-sweep
status
candidate

capabilities

draft launch

accelerator backend
nvidia
arguments
-hf, GestaltLabs/Ornstein3.6-27B-MTP-NSC-ACE-SABER-GGUF:Q4_K_M, --n-gpu-layers, 999, --host, 0.0.0.0, --port, 8080, -c, 262144
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
buun-llama-cpp (86-real, MSVC 19.31, CUDA 12.8)

serving

max concurrency
1
max context tokens
262,144
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
2026-08-31T23:03:15Z
serving.max concurrency · reason server-capacity-derived-from-source-evidence
serving.max concurrency · state known

metadata

localmaxxing · backend
cuda
localmaxxing · batch size
256
localmaxxing · hardware label
RTX 3090
localmaxxing · notes
Batch/ubatch sweep | batch=256 ubatch=128 | MTP n_max=2 p_min=0.75 | KV K=turbo3_tcq V=turbo2_tcq | --fit off --flash-attn on
localmaxxing · observed command
llama-server --threads 12 --prio 3 --n-predict 32768 --batch-size 256 --ubatch-size 128 --flash-attn on --cache-type-k turbo3_tcq --cache-type-v turbo2_tcq --no-mmap --n-gpu-layers all --fit off --ctx-size 262144 --model Ornstein3.6-27B-MTP-NSC-ACE-SABER-Q4_K_M-MTP.gguf --cache-type-k-draft turbo3_tcq --cache-type-v-draft turbo2_tcq --temperature 0.6 --top-p 0.95 --min-p 0.01 --repeat-penalty 1.0 --presence-penalty 0.0 --ctx-checkpoints 8 --cache-ram 8192 --kv-unified --parallel 1 --mmproj mmproj-Ornstein3.6-27B-MTP-NSC-ACE-SABER-F16.gguf --no-mmproj-offload --image-min-tokens 1024 --alias Qwen3.6-27B --props --reasoning on --chat-template-kwargs '{"preserve_thinking":true}' --reasoning-budget 8192 --reasoning-budget-message '. My reasoning budget is exhausted, but I have enough information to answer directly now.' --spec-type draft-mtp --spec-draft-n-max 2 --spec-draft-p-min 0.75
localmaxxing · run id
cmpcxbhcw00fynw017wgjyvsp
localmaxxing · tokenized · arguments
llama-server, --threads, 12, --prio, 3, --n-predict, 32768, --batch-size, 256, --ubatch-size, 128, --flash-attn, on, --cache-type-k, turbo3_tcq, --cache-type-v, turbo2_tcq, --no-mmap, --n-gpu-layers, all, --fit, off, --ctx-size, 262144, --model, Ornstein3.6-27B-MTP-NSC-ACE-SABER-Q4_K_M-MTP.gguf, --cache-type-k-draft, turbo3_tcq, --cache-type-v-draft, turbo2_tcq, --temperature, 0.6, --top-p, 0.95, --min-p, 0.01, --repeat-penalty, 1.0, --presence-penalty, 0.0, --ctx-checkpoints, 8, --cache-ram, 8192, --kv-unified, --parallel, 1, --mmproj, mmproj-Ornstein3.6-27B-MTP-NSC-ACE-SABER-F16.gguf, --no-mmproj-offload, --image-min-tokens, 1024, --alias, Qwen3.6-27B, --props, --reasoning, on, --chat-template-kwargs, {"preserve_thinking":true}, --reasoning-budget, 8192, --reasoning-budget-message, . My reasoning budget is exhausted, but I have enough information to answer directly now., --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/cmpcxbhcw00fynw017wgjyvsp ↗