Use the filters below to narrow your results.
page:
model:
(Click arrow to add another value)
DeepSeek-V4.1-Flash-1017 ·
Qwen3.8-Flash-Next ·
GLM-5.3-Flash (320B MoE, 18B active) ·
DeepSeek-V4-Flash-0731 ·
Qwen3.8-Flash-Next (125B MoE per the recipe) ·
DeepSeek-V4.1-Flash ·
DeepSeek-V4.1-Flash-1017 (748B in the Maya README) ·
DeepSeek-V4.1-Flash-1017 heretic (pqhaz abliteration) ·
GLM-5.3 (full 78-layer DSA MoE, not Flash) ·
GLM-5.3 (full 78-layer, not Flash) ·
GLM-5.3-Flash ·
GLM-5.3-Flash (320B MoE, 18B active, multimodal) ·
GLM-5.3-Flash (321B in the Maya README) ·
MiMo-V2.6-Flash-RL (309B MoE, 15B active; text, image, video and audio input) ·
Qwen/Qwen3.8-Flash-Next-FP8 (revision bcd9f01d) ·
Qwen3.8-27B (dense) ·
Qwen3.8-Flash-Next (nvidia/Qwen3.8-Flash-Next-NVFP4, revision fc694b54, 11 shards, 132.68 GB) ·
Qwen3.8-Flash-Next-Uncensored (refusal-removed, from orcarouter) ·
Other ·
None
quant:
engine:
engine version:
cards:
card count:
topology:
(Click arrow to add another value)
single card ·
engine-managed expert tiering across VRAM, host RAM and NVMe ·
engine-managed split with expert tiering across VRAM, RAM and NVMe ·
layer split (no tensor parallel), no P2P ·
PP2 (layer partition 25,23), no P2P ·
PP2 with PLE CPU offload ·
PP2, MTP 3, no P2P ·
PP2, MTP depth 3 ·
PP3 ·
PP3 (VLLM_PP_LAYER_PARTITION=15,15,13) with DSpark, 5 draft tokens; PCIe Gen2 x4, no P2P ·
PP3, no P2P (works at PP2 on two cards) ·
PP4 (VLLM_PP_LAYER_PARTITION=12,12,12,7) with DSpark, 5 draft tokens; PCIe Gen2 x4, no P2P ·
PP4 (VLLM_PP_LAYER_PARTITION=13,11,11,10), DFlash2 k=7; PCIe Gen2 x4, no P2P ·
PP4, VLLM_PP_LAYER_PARTITION=12,12,12,12, MTP 3, FULL_AND_PIECEWISE CUDA graphs; PCIe Gen2 x4 per card, no P2P. Three cards: PP3 with 16,17,15 or 16,16,16 ·
PP5 with an even 8,8,8,8,8 layer split; the repo's cards are on PCIe Gen2 x16 with no NVLink ·
PP8 (also PP9 and PP10), TP1 ·
PP8 (VLLM_PP_LAYER_PARTITION=7,5,7,5,7,5,7,2) with DFlash2, 7 draft tokens; PCIe Gen2 x4, no P2P; host 2x Xeon E5-2690 v4, 96 GB DDR4 ·
PP8, partition 5,5,5,5,5,5,5,5 ·
PP8, TP1 ·
TEP2 (TP2 plus expert parallel), MTP 6; one card at PCIe Gen2 x16 (x16 mod), one at Gen2 x4; GPU P2P required ·
TP2 over NCCL (GPU P2P), DFlash2 K7 plus hybrid drafting ·
TP2 with expert parallel ·
TP2 with whole-expert expert parallel ·
TP2 x PP3, partition 14,14,12 ·
TP4 ·
TP4 (default) or PP4 ·
TP4 (default) or PP4 for x1/x4 links ·
TP4 (P2P on); expert parallel off ·
TP4 with expert parallel (96 whole experts per card) ·
TP4 with expert parallel, MTP 3 ·
TP4 x PP2, partition 20,20 ·
TP4, P2P on ·
TP4, peer access off (NCCL_P2P_DISABLE=1), host all-reduce ·
TP5 across four CMP 170HX and one RTX PRO 6000 Blackwell (native backend, reduces through host memory); MTP draft on the PRO ·
TP6 (virtual-head padding), PP1 ·
Other ·
None
context:
decode tps:
prefill tps:
vram per card:
power cap:
driver:
unlock build:
status:
tested by:
tested on:
source:
(Click arrow to add another value)
Showing below up to 2 results in range #1 to #2.
& topology:
Other