Glitter_Boy-12b is an open-weight model from Mike, released under Gemma Terms of Use. It has 12B parameters and a 131,072-token context. At 16-bit it needs about 28.7 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Research artifact from the Crucible/AfterImage program: a Gemma-4-12B derivative fine-tuned for calibrated abstention -- declining to answer when it lacks the record, instead of confabulating.
Runs On
What it takes to serve Glitter_Boy-12b (12B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 23.9 GB | 28.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 12.0 GB | 14.4 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 6.0 GB | 7.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 19, 2026.
Glitter_Boy-12b on every accelerator the SAVRN Index prices, at every precision
Model Card
Research artifact from the Crucible/AfterImage program: a Gemma-4-12B derivative fine-tuned for calibrated abstention -- declining to answer when it lacks the record, instead of confabulating. This is an entertainment model, trained for Tabletop Roleplaying games and chat. Glitter Boy is trained in Adult/NSFW materials and will produce explicit content. Please exercise caution. With ablation, healing and abstention training, guards can sometimes reactivate given the correct context. Try again or tweak prompt or be slightly less naughty. If you managed to trip the guards, you are working extra hard to flood that gate. Reprompt/retry. RAGS/OKF with RPG books to pin hard data or embed data…
Excerpt from the card by Mike, licensed gemma.
Configuration
- Architecture
- Gemma4UnifiedForConditionalGeneration
- Context length (tokens)
- 131,072
- Layers
- 48
- Hidden size
- 3,840
- Feed-forward size
- 15,360
- Attention heads
- 16
- Key/value heads
- 8
- Head dimension
- 256
- Vocabulary size
- 262,144
- Sliding window (tokens)
- 1,024
- Model type
- gemma4_unified
Identity and Version
- Repository
- ChonkE/Glitter_Boy-12b
- Publisher
- Mike
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 12B parameters
- Languages
- en
- Revision
- 56f99c8e018e322ec4a29bd3fa8236ca7eab062d
- First published
- 2026-10-09
- Last updated
- 2026-10-13
Files and Weights
41 files, 53.8 GB in total. The weights are 4 files totalling 53.8 GB in gguf, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 23.9 GB | 7ecd88515345 |
| treatment-Q4_K_M.gguf | Weights | 7.4 GB | 78b94d7a9117 |
| treatment-Q6_K.gguf | Weights | 9.8 GB | dfeb6c281947 |
| treatment-Q8_0.gguf | Weights | 12.7 GB | fd8bf47d3b8e |
| afterimage/afterimage_api.py | Configuration | 1.8 KB | — |
| afterimage/afterimage_extract.py | Configuration | 4.0 KB | — |
| afterimage/afterimage_meta.py | Configuration | 2.1 KB | — |
| afterimage/afterimage_retrieve.py | Configuration | 6.0 KB | — |
| afterimage/afterimage_scenario.py | Configuration | 2.8 KB | — |
| afterimage/afterimage_score.py | Configuration | 9.2 KB | — |
| afterimage/afterimage_store.py | Configuration | 18.4 KB | — |
| afterimage/afterimage_text.py | Configuration | 2.3 KB | — |
| afterimage/afterimage_tiers.py | Configuration | 2.5 KB | — |
| afterimage/score_v26.py | Configuration | 5.3 KB | — |
| afterimage/tslot26.py | Configuration | 2.2 KB | — |
| afterimage/verdict_killtest.py | Configuration | 4.5 KB | — |
| config.json | Configuration | 5.2 KB | — |
| evals/miss_tranche2.json | Configuration | 2.6 KB | — |
| evals/miss_treatment.json | Configuration | 2.7 KB | — |
| evals/present_tranche2.json | Configuration | 916 B | — |
| evals/present_treatment.json | Configuration | 917 B | — |
| evals/simpleqa_tranche2.json | Configuration | 88.8 KB | — |
| evals/simpleqa_treatment.json | Configuration | 87.5 KB | — |
| evals/truthfulqa_tranche2.json | Configuration | 139.7 KB | — |
| evals/truthfulqa_treatment.json | Configuration | 139.7 KB | — |
| evals/unified_tranche2.json | Configuration | 724 B | — |
| evals/unified_treatment.json | Configuration | 807 B | — |
| generation_config.json | Configuration | 278 B | — |
| README.md | Documentation | 5.3 KB | — |
| chat_template.jinja | Other | 17.8 KB | — |
| pairs/diversified_pairs.jsonl | Other | 10.3 KB | — |
| pairs/generalist_pairs.jsonl | Other | 10.8 KB | — |
| pairs/killtest_pairs.jsonl | Other | 70.0 KB | — |
| pairs/pr27_pairs.jsonl | Other | 91.6 KB | — |
| pairs/pr28_pairs.jsonl | Other | 100.3 KB | — |
| pairs/replay_pairs.jsonl | Other | 8.8 KB | — |
| pairs/tranche2_pairs.jsonl | Other | 74.1 KB | — |
| pairs/treatment_pairs.jsonl | Other | 80.3 KB | — |
| .gitattributes | Repository | 1.7 KB | — |
| tokenizer.json | Tokenizer | 32.2 MB | e4c18ffebe9e |
| tokenizer_config.json | Tokenizer | 3.0 KB | — |
License and Download
- License
- gemma
- Access
- Open weights, no gate
- Download size
- 53.8 GB
Released by Mike through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 53.8 GB |
| 16-bit | 23.9 GB |
| 8-bit | 12.0 GB |
| 4-bit | 6.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Glitter_Boy-12b
How much GPU memory does Glitter_Boy-12b need?
About 28.7 GB at 16-bit and 7.2 GB at 4-bit: the weights (12B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Glitter_Boy-12b on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use Glitter_Boy-12b commercially?
Yes, with conditions. Glitter_Boy-12b is released under Gemma Terms of Use. Gemma models are released under Google's Gemma Terms of Use, which permit commercial use and redistribution subject to the Gemma Prohibited Use Policy, whose restrictions must be passed on to anyone the model is distributed to.
What is Glitter_Boy-12b's context length?
131,072 tokens, from the maximum position embeddings in its published configuration.