Local AI Registry

Recipe

qwen3-6-35b-a3b-q4-k-m-apple-m5-32gb-llama-cpp-tp1

qwen3-6-35b-a3b-q4-k-m-apple-m5-32gb-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
b9940
Accelerators
1
Tensor parallel
1
Context tokens
64,000
Max concurrency
1
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/unsloth/Qwen3.6-35B-A3B-GGUF
Repository
unsloth/Qwen3.6-35B-A3B-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/cmrdkfqku00mro40154dj4gqc

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-cli
  2. --model
  3. $HOME/models/GGUF/Qwen3.6-35B-A3B/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf
  4. --jinja
  5. --prompt
  6. Write a concise explanation of how transformer attention works.
  7. --n-predict
  8. 512
  9. --temp
  10. 0.6
  11. --top-p
  12. 0.95
  13. --top-k
  14. 20
  15. --min-p
  16. 0.00
  17. --presence-penalty
  18. 0.0
  19. --repeat-penalty
  20. 1.0
  21. --flash-attn
  22. on
  23. --ctx-size
  24. 64000
  25. --mlock
  26. -ngl
  27. 99
  28. -np
  29. 1
FlagValue
--model$HOME/models/GGUF/Qwen3.6-35B-A3B/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf
--promptWrite a concise explanation of how transformer attention works.
--n-predict512
--temp0.6
--top-p0.95
--top-k20
--min-p0.00
--presence-penalty0.0
--repeat-penalty1.0
--flash-attnon
--ctx-size64000
-ngl99
-np1

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
164,000103.534.3—observedqwen3-6-35b-a3b-q4-k-m-apple-m5-32gb-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
apple-m5-32gb
id
qwen3-6-35b-a3b-q4-k-m-apple-m5-32gb-llama-cpp-tp1
model instance id
unsloth-qwen3-6-35b-a3b-gguf--q4-k-m
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-35b-a3b-q4-k-m-apple-m5-32gb-llama-cpp-tp1-sweep
status
candidate

capabilities

engine

name
llama.cpp
version
b9940

serving

max concurrency
1
max context tokens
64,000
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
metal
localmaxxing · hardware label
M5 M5
localmaxxing · observed command
llama-cli \ --model "$HOME/models/GGUF/Qwen3.6-35B-A3B/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf" \ --jinja \ --prompt "Write a concise explanation of how transformer attention works." \ --n-predict 512 \ --temp "0.6" \ --top-p "0.95" \ --top-k "20" \ --min-p "0.00" \ --presence-penalty "0.0" \ --repeat-penalty "1.0" \ --flash-attn on \ --ctx-size "64000" \ --mlock \ -ngl 99 \ -np 1
localmaxxing · run id
cmrdkfqku00mro40154dj4gqc
localmaxxing · tokenized · arguments
llama-cli, --model, $HOME/models/GGUF/Qwen3.6-35B-A3B/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf, --jinja, --prompt, Write a concise explanation of how transformer attention works., --n-predict, 512, --temp, 0.6, --top-p, 0.95, --top-k, 20, --min-p, 0.00, --presence-penalty, 0.0, --repeat-penalty, 1.0, --flash-attn, on, --ctx-size, 64000, --mlock, -ngl, 99, -np, 1
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/cmrdkfqku00mro40154dj4gqc ↗