Local AI Registry

Recipe

qwen3-6-35b-a3b-iq3-xxs-rtx-5060-ti-16gb-llama-cpp-tp1

qwen3-6-35b-a3b-iq3-xxs-rtx-5060-ti-16gb-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
c7896fcf923aabb8fc455ecda891d59e99c8addd
Accelerators
1
Tensor parallel
1
Context tokens
150,000
Max concurrency
1
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

Source
https://www.localmaxxing.com/en/runs/cmsezjk1f00oipp01b6jwneqd

Observed source tokens

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

  1. llamacpp
  2. --server
  3. --host
  4. 0.0.0.0
  5. --port
  6. 8080
  7. --jinja
  8. --n-gpu-layers
  9. 99
  10. -lv
  11. 4
  12. --log-timestamps
  13. --log-colors
  14. on
  15. --checkpoint-min-step
  16. 4000
  17. --ctx-checkpoints
  18. 32
  19. --cache-prompt
  20. --cache-ram
  21. -1
  22. --n-cpu-moe
  23. 0
  24. --temp
  25. 0.7
  26. --top-p
  27. 0.95
  28. --top-k
  29. 20
  30. --min-p
  31. 0.01
  32. --ctx-size
  33. 150000
  34. -b
  35. 2048
  36. -ub
  37. 512
  38. --parallel
  39. 1
  40. --cache-idle-slots
  41. --slot-save-path
  42. /cache/slots/
  43. --flash-attn
  44. on
  45. --keep
  46. 20000
  47. --context-shift
  48. -ctk
  49. q8_0
  50. -ctv
  51. q6_0
  52. --kv-unified
  53. --spec-draft-type-k
  54. turbo4
  55. --spec-draft-type-v
  56. turbo4
  57. --spec-draft-ngl
  58. 99
  59. --spec-draft-n-cpu-moe
  60. 0
  61. --spec-type
  62. draft-mtp,
  63. ngram-mod
  64. --spec-draft-n-max
  65. 2
  66. --spec-draft-ctx-size
  67. 8192
  68. --spec-ngram-mod-n-match
  69. 16
  70. --spec-ngram-mod-n-min
  71. 8
  72. --spec-ngram-mod-n-max
  73. 24
FlagValue
--host0.0.0.0
--port8080
--n-gpu-layers99
-lv4
--log-colorson
--checkpoint-min-step4000
--ctx-checkpoints32
--n-cpu-moe0
--temp0.7
--top-p0.95
--top-k20
--min-p0.01
--ctx-size150000
-b2048
-ub512
--parallel1
--slot-save-path/cache/slots/
--flash-attnon
--keep20000
-ctkq8_0
-ctvq6_0
--spec-draft-type-kturbo4
--spec-draft-type-vturbo4
--spec-draft-ngl99
--spec-draft-n-cpu-moe0
--spec-typedraft-mtp,
--spec-draft-n-max2
--spec-draft-ctx-size8192
--spec-ngram-mod-n-match16
--spec-ngram-mod-n-min8
--spec-ngram-mod-n-max24

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
1150,00085122.3—observedqwen3-6-35b-a3b-iq3-xxs-rtx-5060-ti-16gb-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-5060-ti-16gb
id
qwen3-6-35b-a3b-iq3-xxs-rtx-5060-ti-16gb-llama-cpp-tp1
model instance id
unsloth-qwen3-6-35b-a3b-mtp-gguf--iq3-xxs
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-35b-a3b-iq3-xxs-rtx-5060-ti-16gb-llama-cpp-tp1-sweep
status
candidate

capabilities

draft launch

accelerator backend
nvidia
arguments
-hf, unsloth/Qwen3.6-35B-A3B-MTP-GGUF:IQ3_XXS, --n-gpu-layers, 999, --host, 0.0.0.0, --port, 8080, -c, 150000
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:b2497f8834f5ecb4e38530f6bf2734b8e0be107f0f0857e259672d1cb85b71c2
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
c7896fcf923aabb8fc455ecda891d59e99c8addd

serving

max concurrency
1
max context tokens
150,000
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

metadata

localmaxxing · backend
cuda
localmaxxing · hardware label
RTX 5060 Ti
localmaxxing · observed command
llamacpp --server --host 0.0.0.0 --port 8080 --jinja --n-gpu-layers 99 -lv 4 --log-timestamps --log-colors on --checkpoint-min-step 4000 --ctx-checkpoints 32 --cache-prompt --cache-ram -1 --n-cpu-moe 0 --temp 0.7 --top-p 0.95 --top-k 20 --min-p 0.01 --ctx-size 150000 -b 2048 -ub 512 --parallel 1 --cache-idle-slots --slot-save-path /cache/slots/ --flash-attn on --keep 20000 --context-shift -ctk q8_0 -ctv q6_0 --kv-unified --spec-draft-type-k turbo4 --spec-draft-type-v turbo4 --spec-draft-ngl 99 --spec-draft-n-cpu-moe 0 --spec-type draft-mtp, ngram-mod --spec-draft-n-max 2 --spec-draft-ctx-size 8192 --spec-ngram-mod-n-match 16 --spec-ngram-mod-n-min 8 --spec-ngram-mod-n-max 24
localmaxxing · run id
cmsezjk1f00oipp01b6jwneqd
localmaxxing · tokenized · arguments
llamacpp, --server, --host, 0.0.0.0, --port, 8080, --jinja, --n-gpu-layers, 99, -lv, 4, --log-timestamps, --log-colors, on, --checkpoint-min-step, 4000, --ctx-checkpoints, 32, --cache-prompt, --cache-ram, -1, --n-cpu-moe, 0, --temp, 0.7, --top-p, 0.95, --top-k, 20, --min-p, 0.01, --ctx-size, 150000, -b, 2048, -ub, 512, --parallel, 1, --cache-idle-slots, --slot-save-path, /cache/slots/, --flash-attn, on, --keep, 20000, --context-shift, -ctk, q8_0, -ctv, q6_0, --kv-unified, --spec-draft-type-k, turbo4, --spec-draft-type-v, turbo4, --spec-draft-ngl, 99, --spec-draft-n-cpu-moe, 0, --spec-type, draft-mtp,, ngram-mod, --spec-draft-n-max, 2, --spec-draft-ctx-size, 8192, --spec-ngram-mod-n-match, 16, --spec-ngram-mod-n-min, 8, --spec-ngram-mod-n-max, 24
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/cmsezjk1f00oipp01b6jwneqd ↗