Drilldown: Recipe

DeepSeek V4.1 Flash on 2x 8 GB (Project Maya) (1) · DeepSeek V4.1 Flash on 4x 8 GB (alanoo fidelity build) (1) · DeepSeek V4.1 Flash on 4x 8 GB plus RTX PRO 6000 (ExLlamaV3 TP5) (1) · DeepSeek V4.1 Flash on 6x 8 GB (TP6) (1) · GLM 5.3 Flash on 2x 8 GB (ExLlamaV3 layer split) (1) · GLM 5.3 Flash on 2x 8 GB (ExLlamaV3 TP2) (1) · GLM 5.3 Flash on 2x 8 GB (Project Maya) (1) · Qwen3.8 Flash Next on 2x 8 GB (vektorprime PP2) (1) · Qwen3.8 Flash Next on 3x 8 GB (clee PP3) (1)

DeepSeek-V4.1-Flash-1017 (3) · DeepSeek-V4.1-Flash-1017 (748B in the Maya README) (1) · GLM-5.3-Flash (320B MoE, 18B active) (2) · GLM-5.3-Flash (321B in the Maya README) (1) · Qwen3.8-Flash-Next (2)

AWQ INT4: cyankiwi/Qwen3.8-Flash-Next-AWQ-INT4; 95 GiB BF16 PLE table in host RAM (FP8 table optional) (1) · EXL3 3.05 bpw: turboderp/GLM-5.3-Flash-exl3 (1) · EXL3 3.05 bpw: turboderp/GLM-5.3-Flash-exl3; DFlash2 drafter converted to EXL3 6 bpw (1) · EXL3 3.5 bpw (LibertAIDAI; routed experts 4 bpw on layers 0-9 and 30-39, 3 bpw on 10-29; attention and shared experts 6 bpw) plus the official Engram shards (1) · Hybrid: all 40 expert layers at 3 bits (LibertAI 2.9 bpw pack with layers 18-22 grafted from the 3.5 bpw pack), official FP8 dense tensors (1) · Maya GGUF (Maya-S 340 GB or Maya-XS 277 GB; Engram tables stay on SSD) (1) · Maya-M GGUF (116 GB), every expert in VRAM; Maya-S (96.5 GB) and Maya-S24 (94.7 GB) also reported (1) · Official checkpoint: dense MXFP8 through an SM80 Marlin W8A16 path, routed experts MXFP4 through W4A16; KV FP8 (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)

Custom SM80 vLLM (patched from public commit 79a7108d9) (1) · ExLlamaV3 (patched, pinned) (1) · ExLlamaV3 1.5.4 (built for SM80) with a small OpenAI-compatible server (1) · ExLlamaV3 fork (anoane/exllamav3-anemone) + TabbyAPI fork (anoane/tabbyAPI) (1) · GNmac-lab recipe image 1.2.0 (vLLM with sm_80 patches) plus three patches (1) · Project Maya (Strata and llama.cpp lineage) (1) · Project Maya (Strata and llama.cpp lineage, EXL3 loading, DFlash2 drafter) (1) · vLLM nightly + PR #53899 (a5530b9) + the repo's patch scripts (1) · vLLM nightly 7d0b4e57 plus the author's pure-Python fork clee/vllm-sm80 (branch q38fn/stack) (1)

None (1) · 0.1.dev1+ga5530b90c.d20260829 (1) · Flun/glm53-flash-cmp170hx-exl3 at 5933112 (1) · GNmac-lab at 2ad5741ab1; image ds41-cmp170hx:1.2.0-fidelity (1) · q38fn/stack on nightly-7d0b4e57 (1) · reports 0.0.0; patches/public-base-to-runtime.patch (109 files) (1) · tag recipe-09 (309ef6d1); TabbyAPI deepseek-v41 (ce250ff) (1) · upstream dev 0f685aa9fd plus 9 patches; repo db01c256 (1) · v1.0.6 (this row) (1)

8 GB · Other · None

2 (5) · 3 (1) · 4 (2) · 6 (1)

engine-managed expert tiering across VRAM, host RAM and NVMe (1) · engine-managed split with expert tiering across VRAM, RAM and NVMe (1) · layer split (no tensor parallel), no P2P (1) · PP2 with PLE CPU offload (1) · PP3, no P2P (works at PP2 on two cards) (1) · TP2 over NCCL (GPU P2P), DFlash2 K7 plus hybrid drafting (1) · TP4, P2P on (1) · TP5 across four CMP 170HX and one RTX PRO 6000 Blackwell (native backend, reduces through host memory); MTP draft on the PRO (1) · TP6 (virtual-head padding), PP1 (1)

None (2) · 1,048,576 (1) · 1,048,576 (YaRN x4) (1) · 1,048,576 FP16 (1) · 262,144 (2) · 392,960 (profile q8_384k) (1) · 392,960 per request; 655,360-token Q8 KV pool shared by two requests (1)

60 - 70 (3) · 70 - 110 (6)

None (4) · 514 (1) · 1,533 (1) · 2,116 (1) · 2,514 (1) · 4,831 (1)

64 GB (2) · 64 GB, --gpu-memory-utilization 0.95 (1) · 64 GiB (64,911 MiB usable), GPU_MEMORY_UTILIZATION 0.98 (1) · 64 GiB, gpu-memory-utilization 0.92 (1) · 64 GiB; weights PP0 39.48 GiB, PP1 45.53 GiB; --gpu-memory-utilization 0.9 (1) · about 58 GiB of 64 GiB (116.6 GiB resident across both) (1) · about 62.9 / 62.3 GiB idle, peaks about 64.4 / 63.0 GiB with vision (1) · gpu_split 58 GB per CMP, 70 GB on the PRO (1)

None (6) · 150 W per CMP (PRO 400 W) (1) · 180 W (author's cards); about 170 W with a modest memory OC in his Discord numbers (1) · 200 W (1)

None (6) · 610.57.04 (1) · NVIDIA open 610.43.02 from a cmpunlocker fork with a P2P patch (1) · NVIDIA open 610.57.04 patched with Morrowmake/cmpunlocker 1f2c498 (profile 8gb) (1)

None (6) · cmpunlocker (64 GiB); P2P-capable driver for the NCCL backend (1) · cmpunlocker fork with P2P (without Morrowmake's TRAP31 patch) (1) · Morrowmake/cmpunlocker 1f2c498; modprobe RmForceEnableGen2=1, RMPcieLinkSpeed=0x1, ForceP2P=0x11 (1)

Proven (2) · Reported (7)

68x0 (1) · alanoo (1) · anoane (1) · c133 (1) · nestaa51 (1) · pcsv. (1) · vektor_prime (1) · wgkbfcp (1) · ZackO2o (1)

August 2026 (1) · October 2026 (7)

https://github.com/alanoo81/deepseek-v41-flash-4x-cmp170hx (1) · https://github.com/alexvnesta/glm53-flash-cmp170hx-tp (1) · https://github.com/anoane/09-deepseek-v41-flash-1017-exl3-tp5-1m (1) · https://github.com/clee/qwen3.8-flash-next-3x170hx (1) · https://github.com/Flun/glm53-flash-cmp170hx-exl3 (1) · https://github.com/mw00/project-maya#how-fast-is-it (2) · https://github.com/vektorprime/qwen38-flash-next-pp2 (1) · https://github.com/wgkbfcp/deepseek-v4.1-flash-1017-cmp170hx-recipe (1)