Recipe

qwen3-5-122b-a10b-ud-q4-k-xl-rtx-3090-24gb-llama-cpp-tp4

qwen3-5-122b-a10b-ud-q4-k-xl-rtx-3090-24gb-llama-cpp-tp4

Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.

Record

Status
candidate
Source
localmaxxing
Engine
llama.cpp
Engine version
unknown
Accelerators
4
Tensor parallel
4
Context tokens
65,536
Max concurrency
1
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/unsloth/Qwen3.5-122B-A10B
Repository
unsloth/Qwen3.5-122B-A10B
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/cmowcvbqf00qzp10175xz4fop

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. /workspace/models/UD-Q4_K_XL/Qwen3.5-122B-A10B-UD-Q4_K_XL-00001-of-00003.gguf
  4. --host
  5. 0.0.0.0
  6. --port
  7. 18080
  8. -ngl
  9. 99
  10. --flash-attn
  11. on
  12. --tensor-split
  13. 25,25,25,25
  14. --ctx-size
  15. 65536
  16. --batch-size
  17. 128
  18. --ubatch-size
  19. 128
  20. --cache-type-k
  21. q8_0
  22. --cache-type-v
  23. q8_0
  24. --parallel
  25. 1
FlagValue
-m/workspace/models/UD-Q4_K_XL/Qwen3.5-122B-A10B-UD-Q4_K_XL-00001-of-00003.gguf
--host0.0.0.0
--port18080
-ngl99
--flash-attnon
--tensor-split25,25,25,25
--ctx-size65536
--batch-size128
--ubatch-size128
--cache-type-kq8_0
--cache-type-vq8_0
--parallel1

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
165,536467.251.3121,716.4observedqwen3-5-122b-a10b-ud-q4-k-xl-rtx-3090-24gb-llama-cpp-tp4-sweep

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

hardware count
4
hardware id
rtx-3090-24gb
id
qwen3-5-122b-a10b-ud-q4-k-xl-rtx-3090-24gb-llama-cpp-tp4
model instance id
unsloth-qwen3-5-122b-a10b--ud-q4-k-xl
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-5-122b-a10b-ud-q4-k-xl-rtx-3090-24gb-llama-cpp-tp4-sweep
status
candidate

capabilities

engine

name
llama.cpp
version
unknown

serving

max concurrency
1
max context tokens
65,536
tensor parallel
4
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 3090
localmaxxing · notes
Benchmark run with llama-benchy (v0.3.7). Single run at 64k context depth. Unsloth GGUF distribution of Qwen3.5-122B-A10B finetune, quantized to UD-Q4_K_XL. 4x RTX 3090 with tensor parallelism.
localmaxxing · observed command
llama-server -m /workspace/models/UD-Q4_K_XL/Qwen3.5-122B-A10B-UD-Q4_K_XL-00001-of-00003.gguf --host 0.0.0.0 --port 18080 -ngl 99 --flash-attn on --tensor-split 25,25,25,25 --ctx-size 65536 --batch-size 128 --ubatch-size 128 --cache-type-k q8_0 --cache-type-v q8_0 --parallel 1
localmaxxing · run id
cmowcvbqf00qzp10175xz4fop
localmaxxing · tokenized · arguments
llama-server, -m, /workspace/models/UD-Q4_K_XL/Qwen3.5-122B-A10B-UD-Q4_K_XL-00001-of-00003.gguf, --host, 0.0.0.0, --port, 18080, -ngl, 99, --flash-attn, on, --tensor-split, 25,25,25,25, --ctx-size, 65536, --batch-size, 128, --ubatch-size, 128, --cache-type-k, q8_0, --cache-type-v, q8_0, --parallel, 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/cmowcvbqf00qzp10175xz4fop