Milo-Ark is the checksummed model store on miloshare. Preservation is the contract. Offline serve is recorded only when a smoke result sits beside the checksums. Storage rebuild lives in the DS1823xs+ post.
Now: every tree in the Ark is checksummed and nothing is downloading. Fourteen trees joined the inventory since September 19, and the 3.77 TB audnai/penclaw-GLM-5.3-abliterated tree passed its NAS-side verify (738/738). The biggest new tree is moonshotai/Kimi-K3.1-Pro (1.56 TB, 98/98 LFS hashes checked on the NAS against the Hub). deepseek-ai/DeepSeek-V4.1-Flash-1017 is the first tree whose manifest records a passed smoke test: vLLM on the GB300 Station, tool gauntlet 12/12, GSM8K-200 96.5%, needle-checksum 500/500, NAS replica 85/85. That serve ran from the Station's NVMe copy of the same revision, so it is not yet an offline-from-the-Ark serve.
Closed today: meta-models/Muse-Spark-Open-816B-A42B-NVFP4 (533.0 GB) was hashed on the NAS and all 128 large files match the Hub's sha256. The stalled abenzerps/Qwen-Image-2.1-Uncensored-GGUF copy was finished from the local stage and matches the Hub 17/17. nvidia/DeepSeek-V4.1-Flash-NVFP4 had checksums but no record of a check on the NAS; sha256sum -c on the NAS now passes 83/83.
Checksummed inventory
| Repo | Rev | Size | n | State |
|---|---|---|---|---|
abenzerps/Qwen-Image-2.1-Uncensored-GGUF new | 6b34e59458d3 | 105.2 GB | 25 | checksum |
audnai/penclaw-GLM-5.3-abliterated new | 027ae98a6f9d | 3.77 TB | 739 | checksum |
avlp12/Kimi-K2.7-Code-Alis-MLX-Dynamic-3.6bpw | 2407e9b7948a | 464.9 GB | 118 | checksum |
Blackfrost-AI/DeepSeek-V4.1-Flash-1017-DERISKED-FP8 new | 4bb7df57c6bf | 510.3 GB | 88 | checksum |
Blackfrost-AI/KIMI-K3-Q2_K-GGUF-ABLITERATED new | b233c5abd8be | 1.01 TB | 51 | checksum |
Blackfrost-Research/GLM-5.3-F.U-AnthraClaud-Edition-BF16 new | eac6aa06d2b1 | 1.51 TB | 169 | checksum · >HW |
Blackfrost-Research/GLM-5.3-Flash-DERISKED-NVFP4 | ac08b5ee7c8b | 205.1 GB | 133 | checksum |
cebeuq/DeepSeek-V4-Flash-0731-abliterated new | 21bd923c2574 | 168.4 GB | 89 | checksum |
dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8 | d61c59ea5e51 | 510.3 GB | 62 | checksum |
dealignai/GLM-5.3-ABLITERATED-NVFP4 | 371bdb985d01 | 464.9 GB | 294 | checksum |
dealignai/GLM-5.3-Flash-ABLITERATED-FP8 | 3eb1880f95b1 | 328.4 GB | 75 | checksum |
deepseek-ai/DeepSeek-V4-Flash | 60d8d70770c6 | 159.6 GB | 73 | checksum |
deepseek-ai/DeepSeek-V4-Flash-0731 | 7872f01b1d1f | 166.9 GB | 78 | checksum |
deepseek-ai/DeepSeek-V4-Flash-DSpark | 913f0657a874 | 166.9 GB | 74 | checksum |
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp | 6821d6ad3681 | 167.8 GB | 84 | checksum · served |
deepseek-ai/DeepSeek-V4-Pro | b5968e9190ef | 864.7 GB | 91 | checksum |
deepseek-ai/DeepSeek-V4.1-Flash | df42c109f1de | 510.3 GB | 88 | checksum · served |
deepseek-ai/DeepSeek-V4.1-Flash-1017 new | 1bae0e108899 | 510.3 GB | 85 | checksum · smoke passed |
drowzeys/keys-DeepSeekV4-Flash-GA-0731-Dspark-Abliterated-32-32 | 6c47916f85e5 | 166.9 GB | 80 | checksum |
drowzeys/keys-DeepSeekV4-Flash-GA-0731-Dspark-Abliterated-Anchored-Tensors new | a1e693795173 | 166.9 GB | 81 | checksum |
drowzeys/keys-DeepSeekV4Flash-Vision-EXP-ablit new | 48095b3452a1 | 167.8 GB | 85 | checksum |
drowzeys/keys-GLM-5.3-Flash-NVFP4-ablit-l15-43-mtp-l45 | 80b6d18d77e3 | 197.9 GB | 34 | checksum |
google/gemma-4-26b-a4b-it | 20da991ab4af | 51.6 GB | 12 | checksum |
incoai/GLM-5.3-DFlash2 | 425aa615ce32 | 4.9 GB | 9 | checksum |
incoai/GLM-5.3-Flash-DFlash2 | 7d74cdd881ed | 2.3 GB | 8 | checksum |
incoai/GLM-5.3-NVFP4 | 54e52520606f | 464.9 GB | 101 | checksum |
incoai/Qwen3.8-27B-DFlash2 | adde41d8fde3 | 3.8 GB | 9 | checksum |
k-l-lambda/kimi-k2.7-code-eagle3-mla | e2b24ca46993 | 4.1 GB | 4 | checksum |
LibertAIDAI/GLM-5.3-Flash-NVFP4 | aa28e1f54130 | 194.7 GB | 134 | checksum |
lmstudio-community/GLM-4.5-Air-MLX-8bit | 251aed6f354a | 113.6 GB | 32 | checksum |
lmstudio-community/MiniMax-M2.5-GGUF | 93cbd1a0bf00 | 430.9 GB | 14 | checksum |
lmstudio-community/MiniMax-M2.5-MLX-4bit | 9af822e97aaa | 128.7 GB | 35 | checksum |
lmstudio-community/Qwen2.5-Coder-32B-Instruct-MLX-4bit | 96572f69fd88 | 18.4 GB | 14 | checksum |
lmstudio-community/Qwen2.5-Coder-32B-Instruct-MLX-8bit | 7bd37ed2a7f0 | 34.8 GB | 17 | checksum |
lmstudio-community/Qwen3.5-397B-A17B-MLX-4bit | 76a66de5151f | 223.9 GB | 58 | checksum |
local-inference-lab/Qwen3.8-Flash-Next-NVFP4-4p89 | ee0cea634a37 | 105.9 GB | 55 | checksum |
malaiwah/GLM-5.3-Flash-TR3-6bpw | 1bdd7d3d9aea | 253.7 GB | 268 | checksum |
meta-models/Muse-Spark-Open-816B-A42B-NVFP4 new | 4239e4db1937 | 533.0 GB | 146 | checksum |
mlx-community/gemma-3-12b-it-qat-4bit | 66fc51ef2577 | 8.1 GB | 15 | checksum |
mlx-community/magnum-v4-72b-4bit | 8c5ef3197ee0 | 40.9 GB | 18 | checksum |
moonshotai/Kimi-K2.7-Code | 74797c9c6237 | 595.2 GB | 87 | checksum |
moonshotai/Kimi-K3 | a590ce090cb0 | 1.56 TB | 122 | checksum · >HW |
moonshotai/Kimi-K3.1-Pro new | 6d0a02029567 | 1.56 TB | 116 | checksum |
NousResearch/Hermes-4-70B | d5dec2bd6b39 | 141.1 GB | 39 | checksum |
nvidia/DeepSeek-V4.1-Flash-NVFP4 | 3431dde3247c | 527.3 GB | 83 | checksum |
nvidia/GLM-5.2-NVFP4 | aec724e8c7b8 | 464.9 GB | 57 | checksum |
nvidia/GLM-5.3-Flash-NVFP4 | 09b04e5e74bc | 204.5 GB | 45 | checksum · served |
nvidia/Kimi-K2.7-Code-NVFP4 | 9f28d60908c9 | 595.2 GB | 87 | checksum |
nvidia/Kimi-K3-NVFP4 | b2428a0b83a8 | 1.61 TB | 120 | checksum · >HW |
nvidia/MiniMax-M2.7-NVFP4 | e79701cb1f9d | 139.9 GB | 30 | checksum |
nvidia/parakeet-tdt-0.6b-v2 | 1b149a358935 | 2.5 GB | 3 | checksum |
nvidia/Qwen3.8-Flash-Next-NVFP4 | fc694b54fb01 | 132.7 GB | 25 | checksum · served |
Qwen/Qwen-Image-2.1 new | b3179ad355be | 33.1 GB | 29 | checksum |
Qwen/Qwen3-Coder-30B-A3B-Instruct | b2cff646eb4b | 61.1 GB | 28 | checksum |
Qwen/Qwen3-Coder-Next | a7fbcb5c0e12 | 159.4 GB | 52 | checksum |
Qwen/Qwen3-Embedding-8B | 1d8ad4ca9b3d | 15.2 GB | 17 | checksum |
Qwen/Qwen3-Reranker-4B | 22e683669bc0 | 8.1 GB | 17 | checksum |
Qwen/Qwen3-VL-30B-A3B-Instruct | 9c4b90e1e4ba | 62.2 GB | 25 | checksum |
Qwen/Qwen3.6-27B | 6a9e13bd6fc8 | 55.6 GB | 29 | checksum |
Qwen/Qwen3.6-35B-A3B | 995ad96eacd9 | 71.9 GB | 40 | checksum |
Qwen/Qwen3.8-27B | 1d4bf0f2ff60 | 55.6 GB | 32 | checksum |
RadixArk/GLM-5.3-NVFP4 new | 6e389189d564 | 464.9 GB | 58 | checksum |
RedHatAI/GLM-5.3-Flash-NVFP4 | 240131d6a447 | 197.9 GB | 23 | checksum |
unsloth/Qwen3.5-35B-A3B-GGUF | bc014a17be43 | 22.0 GB | 5 | checksum |
XiaomiMiMo/MiMo-V2.6-Pro-RL new | 54b10491b181 | 573.5 GB | 158 | checksum |
zai-org/GLM-5.3 new | ba32f8c19c87 | 31.6 MB | 15 | checksum · delta |
zai-org/GLM-5.3 | e0b07fd2751b | 755.7 GB | 158 | checksum |
zai-org/GLM-5.3-BF16 | 304b8051cfb2 | 1.51 TB | 295 | checksum · >HW |
zai-org/GLM-5.3-Flash | 04c4e9e95c5d | 328.4 GB | 76 | checksum |
zai-org/GLM-5.3-Flash-BF16 | f12e0fe1f6b2 | 642.7 GB | 134 | checksum |
new marks trees added or verified since September 19. served means a GB300 serving note in that tree's manifest (Vision-Exp, V4.1-Flash, nvidia GLM-5.3-Flash-NVFP4, nvidia Qwen3.8-Flash-Next-NVFP4). smoke passed means the manifest records a passed smoke test. >HW means runtime proof is N/A because the snapshot exceeds owned hardware. delta is the zai-org/GLM-5.3 chat-template revision: its 141 weight files are byte-identical to e0b07fd2751b, so only the changed small files are stored. Sizes are bytes in original/; n is SHA256SUMS lines. An M3 Ultra parking lot under M3UltraModels/ is not in this table.
Quarantine
| Repo | Why |
|---|---|
inferencerlabs/Kimi-K2.6-MLX-3.6bit | source gone; shards 44/48 and 47/48 missing |
DevQuasar/moonshotai.Kimi-K2.5-GGUF | only shard 40/42; size mismatch |
lmstudio-community/gemma-4-26B-A4B-it-GGUF | three GGUF sizes ≠ pinned revision |
Done when: pinned files + SHA256 + provenance, then a local-files-only serve and smoke test stored beside the checksums. Checksum ≠ runtime proof. Serving notes ≠ archive smoke files. A smoke test on a staged copy ≠ a serve from the Ark itself. SHA256SUMS written ≠ sha256sum -c passed.