Drilldown: Recipe

DeepSeek V4.1 Flash on 4x 8 GB (CurtisAccelerate TC-MoE) (1) · GLM 5.3 Flash on 2x 8 GB (Project Maya) (1) · GLM 5.3 on 8x 8 GB (vllm-170hx-glm5) (1)

DeepSeek-V4.1-Flash-1017 (1) · GLM-5.3 (full 78-layer DSA MoE, not Flash) (1) · GLM-5.3-Flash (321B in the Maya README) (1)

GLM-5.3-Int4-Int8Mix-AWQ-g64; drafter incoai/GLM-5.3-DFlash2 (1) · Hybrid: LibertAI EXL3 2.9 bpw routed experts (layers 0-39), official FP8/FP4 for attention, shared experts, head, draft and Engram; KV fp8_ds_mla (1) · Maya-M GGUF (116 GB), every expert in VRAM; Maya-S (96.5 GB) and Maya-S24 (94.7 GB) also reported (1)

Morrowmake/vllm-cmp170hx (bind-mounted sm80 package) plus the author's tensor-core MoE kernels (1) · Morrowmake/vllm-cmp170hx branch ampere-glm53 plus fail-closed source patches (1) · Project Maya (Strata and llama.cpp lineage, EXL3 loading, DFlash2 drafter) (1)

v1.0.6 (this row) (1) · vLLM 0.30.1rc1, fork commit 378c37b0 (1) · vllm-cmp170hx 18305948ac; TC-MoE fb9786ec (1)

8 GB (3)

2 (1) · 4 (1) · 8 (1)

engine-managed split with expert tiering across VRAM, RAM and NVMe (1) · PP8 (also PP9 and PP10), TP1 (1) · TP4, peer access off (NCCL_P2P_DISABLE=1), host all-reduce (1)

None (1) · 1,048,576 per request (1) · 262,144 (weights address 1,048,576) (1)

48.7 (1) · 61.3 (1) · 144.2 (1)

514 (1) · 1,523 (1) · 1,833 (1)

64 GB (1) · 64 GiB, GPU_UTIL 0.96 at PP8 (1) · about 62.5 GB used of 64 GB, --gpu-memory-utilization 0.97 (1)

None (1) · 140 W (1) · 170 W (author's 10-card run) (1)

None · 580 or newer (measured on 610.57.04) · NVIDIA 580 or newer; P2P needs the Morrowmake/cmpunlocker build · 580 or newer (measured on 610.57.04); open modules with HMM for the table-on-disk mode · 610.43.03 · 610.57.04 · CUDA 13 capable; measured on 610.43.02 · NVIDIA 610.43.02 · NVIDIA open 610.43.02 from a cmpunlocker fork with a P2P patch · NVIDIA open 610.57.04 patched with Morrowmake/cmpunlocker 1f2c498 (profile 8gb) · nvidia-open 610.43.02 + cmpunlocker v0.5 or newer · Other

None (3)

Proven (1) · Reported (2)

CurtisAccelerate (1) · reala7096 (1) · ZackO2o (1)

September 29, 2026 (1) · October 22, 2026 (1)

https://github.com/arvkmr/vllm-170hx-glm5 (1) · https://github.com/CurtisAccelerate/DeepSeekv41-Flash-1017-EXL3-CMP170HX (1) · https://github.com/mw00/project-maya#how-fast-is-it (1)