Page values for "MiMo V2.6 Flash on 4x 8 GB"
From CMP Wiki
More actions
"Recipe" values
1 row is stored for this page| Field | Field type | Allowed values | Value |
|---|---|---|---|
| page | Page | MiMo V2.6 Flash on 4x 8 GB | |
| model | String | MiMo-V2.6-Flash-RL (309B MoE, 15B active; text, image, video and audio input) | |
| quant | String | NVFP4 (ModelOpt mixed precision): ProCreations/MiMo-V2.6-Flash-RL-NVFP4, run as 4-bit weights with BF16 activations through Marlin | |
| engine | String | Morrowmake/vllm-cmp170hx image (tag qwen38-recipe-1.0.0) plus one mounted file (vLLM #58235 vision-sink fix) | |
| engine_version | String | recipe 0.2.1; engine 18305948ac | |
| cards | String | 8 GB · 10 GB · both | 8 GB |
| card_count | Integer | 4 | |
| topology | String | TP4 (P2P on); expert parallel off | |
| context | String | 1,048,576 | |
| decode_tps | Float | 274.9 | |
| prefill_tps | Float | 2,722.0 | |
| vram_per_card | String | 64 GiB; weights 45.5 GiB plus 10 GiB KV per card; GPU_UTIL 0.94 (fixed) | |
| power_cap | String | 180 W | |
| driver | String | 580 or newer (measured on 610.57.04) | |
| unlock_build | String | cmpunlocker (64 GiB, 74 SMs) with a P2P-capable build | |
| flags | Text | --tensor-parallel-size 4 --kv-cache-dtype bfloat16 --block-size 128 --max-model-len 1048576 --gpu-memory-utilization 0.94 --kv-cache-memory 10737418240 --max-num-seqs 32 --max-num-batched-tokens 2048; DFlash 7/5/3 by load | |
| status | String | Proven · Reported · Theory · Contested · Warning | Reported |
| tested_by | String | nathang6866 | |
| tested_on | Date | 2026-10-18 | |
| source | String | https://github.com/ElectroHobo/mimo-v26-flash-cmp170hx-recipe | |
| files | Wikitext |