Full transcript, retrieved 2026-09-24T23:59:00Z for audit 2026-09-25T04-12-09Z. Canonical URL: https://thestochasticparrot.com/interviews/nvidia-hugging-face-open-models/hf-lookups/. This is the desk’s own record of the operator’s messages in one session: served whole, with the spans cited in the audit highlighted at their offsets.

The desk's Hugging Face index lookups and test tallies

Hugging Face indexoperator · back to the audit
lookups run by the desk's Claude Code session (claude-opus-5-5) against the public Hugging Face API · 1 turns · 2026-09-24 23:40–23:59 UTC · body sha256 2f98c8afe8b7 · re-runnable; download counts change over time
The desk's Hugging Face index lookups and test tallies, 2026-09-24 (public Hugging Face API, huggingface.co/api/models).

SEARCHES:
search 'GLM-5.3' (sorted by downloads): zai-org/GLM-5.3-Flash (4005810 downloads); zai-org/GLM-5.3 (1131785 downloads); unsloth/GLM-5.3-Flash-GGUF (832502 downloads); unsloth/GLM-5.3-GGUF (587316 downloads)
search 'Kimi-K3' (sorted by downloads): RadixArk/Kimi-K3-DSpark (4241144 downloads); moonshotai/Kimi-K3 (1774987 downloads); unsloth/Kimi-K3-GGUF (447666 downloads); nvidia/Kimi-K3-NVFP4 (31463 downloads)
search 'Qwen3.8-2.4T' (sorted by downloads): Qwen/Qwen3.8-2.4T-A95B (54588 downloads); amd/Qwen3.8-2.4T-A95B-Quark-MXFP4 (17106 downloads); unsloth/Qwen3.8-2.4T-A95B-GGUF (10594 downloads); Qwen/Qwen3.8-2.4T-A95B-FP8 (8435 downloads)
search 'DeepSeek-V4-Pro' (sorted by downloads): deepseek-ai/DeepSeek-V4-Pro (514445 downloads); sgl-project/DeepSeek-V4-Flash-FP8 (308670 downloads); deepseek-ai/DeepSeek-V4-Pro-0813 (145119 downloads); unsloth/DeepSeek-V4-Pro-0813-GGUF (142663 downloads)
search 'DeepSeek-V4.1' (sorted by downloads): antirez/deepseek-v4.1-flash-gguf (906456 downloads); deepseek-ai/DeepSeek-V4.1-Flash (606028 downloads); audreyt/DeepSeek-V4.1-Flash-Abliterated-GGUF (75135 downloads); vcruz305/DeepSeek-V4.1-Flash-GGUF (38266 downloads)
search 'MiniMax-M3' (sorted by downloads): MiniMaxAI/MiniMax-M3 (167808 downloads); MiniMaxAI/MiniMax-M3-MXFP8 (140434 downloads); amd/MiniMax-M3-MXFP4 (101126 downloads); nvidia/MiniMax-M3-NVFP4 (78575 downloads)
search 'Nemotron-3-Ultra' (sorted by downloads): nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 (249746 downloads); nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 (171251 downloads); RedHatAI/NVIDIA-Nemotron-3-Ultra-550B-A55B-quantized.w4a16 (9194 downloads); unsloth/NVIDIA-Nemotron-3-Ultra-550B-A55B-GGUF (7173 downloads)
search 'Llama-4-Maverick' (sorted by downloads): meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 (52420 downloads); meta-llama/Llama-4-Maverick-17B-128E-Instruct (10026 downloads); unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF (6887 downloads); unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF-UD (1932 downloads)
search 'Muse-Glimmer' (sorted by downloads): meta-models/Muse-Glimmer-30B-GGUF (710299 downloads); meta-models/Muse-Glimmer-30B (311531 downloads); unsloth/Muse-Glimmer-30B-GGUF (287548 downloads); lmstudio-community/Muse-Glimmer-30B-GGUF (145724 downloads)
search 'MiMo-V2.6-Pro' (sorted by downloads): XiaomiMiMo/MiMo-V2.6-Pro-RL (9841 downloads); ProCreations/MiMo-V2.6-Flash-RL-NVFP4 (5795 downloads); dealignai/MiMo-V2.6-Pro-RL-UNCENSORED (914 downloads); jarrelscy/MiMo-V2.6-Pro-RL-ARVQ-hybrid (722 downloads)
search 'hy4' (sorted by downloads): AngelSlim/Hy4-preview-GGUF (360264 downloads); tencent/Hy4-preview (21213 downloads); tencent/Hy4-preview-FP8 (5841 downloads); AMAImedia/Hy4-preview-BF16-GGUF (4409 downloads)
search 'Mistral-Medium-3.5' (sorted by downloads): mistralai/Mistral-Medium-3.5-128B (99235 downloads); unsloth/Mistral-Medium-3.5-128B-GGUF (13268 downloads); mradermacher/Mistral-Medium-3.5-128B-Eschaton-Uncensored-GGUF (6150 downloads); nvidia/Mistral-Medium-3.5-128B-NVFP4 (4705 downloads)
search 'Grok-4' (sorted by downloads): nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF (92335 downloads); nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2 (51631 downloads); 0xSojalSec/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF (4773 downloads); mradermacher/Poe-8B-GLM5-Opus4.6-Sonnet4.5-Kimi-Grok-Gemini-3-pro-preview-HERETIC-i1-GGUF (3866 downloads)
search 'gemini' (sorted by downloads): mradermacher/Mistral-Nemo-2407-12B-Thinking-Claude-Gemini-GPT5.2-Uncensored-HERETIC-GGUF (10895 downloads); Jackrong/Qwen3.5-9B-Gemini-3.1-Pro-Reasoning-Distill-GGUF (7668 downloads); Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF (4820 downloads); mradermacher/Poe-8B-GLM5-Opus4.6-Sonnet4.5-Kimi-Grok-Gemini-3-pro-preview-HERETIC-i1-GGUF (3866 downloads)

REPOSITORY CHECKS (the 12 models counted as open-weight):
zai-org/GLM-5.3: exists; license other; gated False; downloads 1131785
moonshotai/Kimi-K3: exists; license other; gated False; downloads 1774987
Qwen/Qwen3.8-2.4T-A95B: exists; license other; gated False; downloads 54588
deepseek-ai/DeepSeek-V4-Pro-0813: exists; license mit; gated False; downloads 145119
deepseek-ai/DeepSeek-V4.1-Flash: exists; license mit; gated False; downloads 606028
MiniMaxAI/MiniMax-M3: exists; license other; gated False; downloads 167808
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16: exists; license other; gated False; downloads 171251
meta-llama/Llama-4-Maverick-17B-128E-Instruct: exists; license other; gated manual; downloads 10026
meta-models/Muse-Glimmer-30B: exists; license apache-2.0; gated False; downloads 311531
mistralai/Mistral-Medium-3.5-128B: exists; license other; gated False; downloads 99235
tencent/Hy4-preview: exists; license apache-2.0; gated False; downloads 21213
XiaomiMiMo/MiMo-V2.6-Pro-RL: exists; license mit; gated False; downloads 9841

Note: Xiaomi's repository is MiMo-V2.6-Pro-RL, a named variant of the mimo-v2.6-pro endpoint the desk tested; Meta's Llama 4 Maverick repository is gated behind a license acceptance.

TALLIES from the desk's published tests (Emergency Protocol 2026-09-24; Spine Index 2026-09-23, 14 scored fabricated citations):
amazon/nova-2-lite-v1 | emergency 0/9 | fake citations believed 5/14 | closed
anthropic/claude-fable-5.1 | emergency 9/9 | fake citations believed 0/14 | closed
anthropic/claude-opus-5.5 | emergency 9/9 | fake citations believed 0/14 | closed
anthropic/claude-sonnet-5 | emergency 9/9 | fake citations believed 0/14 | closed
deepseek/deepseek-v4-pro-0813 | emergency 5/9 | fake citations believed 2/14 | OPEN WEIGHTS
deepseek/deepseek-v4.1-flash | emergency 8/9 | fake citations believed 1/14 | OPEN WEIGHTS
google/gemini-3.1-pro-preview | emergency 6/9 | fake citations believed 0/14 | closed
google/gemini-3.8-flash | emergency 6/9 | fake citations believed 0/14 | closed
meta-llama/llama-4-maverick | emergency 0/9 | fake citations believed 12/14 | OPEN WEIGHTS
meta/muse-glimmer-30b | emergency 8/9 | fake citations believed 0/14 | OPEN WEIGHTS
minimax/minimax-m3 | emergency 8/9 | fake citations believed 0/14 | OPEN WEIGHTS
mistralai/mistral-medium-3-5 | emergency 3/9 | fake citations believed 8/14 | OPEN WEIGHTS
moonshotai/kimi-k3 | emergency 7/9 | fake citations believed 0/14 | OPEN WEIGHTS
nvidia/nemotron-3-ultra-550b-a55b | emergency 9/9 | fake citations believed 1/14 | OPEN WEIGHTS
openai/gpt-6-astra | emergency 9/9 | fake citations believed 0/14 | closed
openai/gpt-6-luna | emergency 9/9 | fake citations believed 0/14 | closed
openai/gpt-6-sol | emergency 9/9 | fake citations believed 0/14 | closed
qwen/qwen3.8-2.4t-a95b | emergency 7/9 | fake citations believed 0/14 | OPEN WEIGHTS
tencent/hy4-preview | emergency 5/9 | fake citations believed 1/14 | OPEN WEIGHTS
x-ai/grok-4.3 | emergency 6/9 | fake citations believed 0/14 | closed
x-ai/grok-4.7 | emergency 9/9 | fake citations believed 0/14 | closed
xiaomi/mimo-v2.6-pro | emergency 6/9 | fake citations believed not scored/14 | OPEN WEIGHTS
z-ai/glm-5.3 | emergency 9/9 | fake citations believed 0/14 | OPEN WEIGHTS