Drilldown: Recipe - CMP Wiki

Jump to content

Choose a table:

Recipe > tested by

: ZackO2o

or c133

or matstatman

Use the filters below to narrow your results.

GLM 5.3 Flash on 2x 8 GB (Project Maya) (1) · Qwen3.8 Flash Next on 3x 8 GB (clee PP3) (1) · Qwen3.8 Flash Next on 4x 8 GB (1)

GLM-5.3-Flash (321B in the Maya README) (1) · Qwen3.8-Flash-Next (2)

AWQ W4A16: wtdcode/Qwen3.8-Flash-Next-AWQ-W4A16 (1) · Maya-M GGUF (116 GB), every expert in VRAM; Maya-S (96.5 GB) and Maya-S24 (94.7 GB) also reported (1) · W4A16 AutoRound: Intel/Qwen3.8-Flash-Next-W4A16-AutoRound; 95.4 GiB BF16 PLE table streamed from NVMe; KV fp8 e4m3 (software-decoded on SM80) (1)

Project Maya (Strata and llama.cpp lineage, EXL3 loading, DFlash2 drafter) (1) · vLLM nightly 7d0b4e57 plus the author's pure-Python fork clee/vllm-sm80 (branch q38fn/stack) (1) · wtdcode/vllm-backport (image lazymio/vllm-backport:latest-sm80) (1)

latest-sm80 tag as of 2026-10-16 (1) · q38fn/stack on nightly-7d0b4e57 (1) · v1.0.6 (this row) (1)

8 GB (3)

2 (1) · 3 (1) · 4 (1)

engine-managed split with expert tiering across VRAM, RAM and NVMe (1) · PP3, no P2P (works at PP2 on two cards) (1) · TP4 with expert parallel, MTP 3 (1)

None (1) · 1,000,000 (YaRN x4 from 262,144) (1) · 1,048,576 (YaRN x4) (1)

61.3 (1) · 74.4 (1) · 116.49 (1)

514 (1) · 2,101.4 (1) · 4,831 (1)

64 GB (2) · 64 GiB, gpu-memory-utilization 0.92 (1)

None (3)

None (3)

None (3)

Proven (1) · Reported (2)

October 10, 2026 (1) · October 16, 2026 (1)

https://discord.com/channels/1518634379604136178/1528826330454429706/1560602001908170753 (1) · https://github.com/clee/qwen3.8-flash-next-3x170hx (1) · https://github.com/mw00/project-maya#how-fast-is-it (1)

Showing below up to 3 results in range #1 to #3.

G

GLM 5.3 Flash on 2x 8 GB (Project Maya)

Q

Qwen3.8 Flash Next on 3x 8 GB (clee PP3)

Qwen3.8 Flash Next on 4x 8 GB

Retrieved from "https://cmp.wiki/wiki/Special:Drilldown/Recipe"