Open-weight model
gemma-4-31b-fve-msm-control-s3
by Joshua Fonseca joshycodes/gemma-4-31b-fve-msm-control-s3
gemma-4-31b-fve-msm-control-s3 is an open-weight model from Joshua Fonseca, released under other. It has 31.3B parameters and a 262,144-token context. At 16-bit it needs about 75.1 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Private research artifact — do not redistribute. Anthropic Fellows project on flourishing-framed character training (Wang & Jermyn pitch, 2026-04-22).
Runs On
What it takes to serve gemma-4-31b-fve-msm-control-s3 (31.3B 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 | 62.5 GB | 75.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 31.3 GB | 37.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 15.6 GB | 18.8 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 20, 2026.
gemma-4-31b-fve-msm-control-s3 on every accelerator the SAVRN Index prices, at every precision
Model Card
Private research artifact — do not redistribute. Anthropic Fellows project on flourishing-framed character training (Wang & Jermyn pitch, 2026-04-22). Full run config in trainrunconfig.json; evaluate with uv run eval.py --model joshycodes/gemma-4-31b-fve-msm-control-s3 --eval all.
Excerpt from the card by Joshua Fonseca, licensed other.
Configuration
- Architecture
- Gemma4ForConditionalGeneration
- Context length (tokens)
- 262,144
- Layers
- 60
- Hidden size
- 5,376
- Feed-forward size
- 21,504
- Attention heads
- 32
- Key/value heads
- 16
- Head dimension
- 256
- Vocabulary size
- 262,144
- Sliding window (tokens)
- 1,024
- Model type
- gemma4
Identity and Version
- Repository
- joshycodes/gemma-4-31b-fve-msm-control-s3
- Publisher
- Joshua Fonseca
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 31.3B parameters
- Languages
- Not stated by the source
- Revision
- acb0b77aa2ff093948fbfb8c1b06e344e29b5559
- First published
- 2026-10-15
- Last updated
- 2026-10-16
Files and Weights
14 files, 62.6 GB in total. The weights are 3 files totalling 62.5 GB in bin, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00002.safetensors | Weights | 49.9 GB | a1f7e0fa0321 |
| model-00002-of-00002.safetensors | Weights | 12.6 GB | dbe7eba96c2c |
| training_args.bin | Weights | 5.6 KB | 05a062ab57bb |
| config.json | Configuration | 5.5 KB | — |
| generation_config.json | Configuration | 194 B | — |
| model.safetensors.index.json | Configuration | 120.2 KB | — |
| processor_config.json | Configuration | 1.7 KB | — |
| train_run_config.json | Configuration | 2.8 KB | — |
| README.md | Documentation | 1.1 KB | — |
| chat_template.jinja | Other | 269 B | — |
| runs/train.log | Other | 15.9 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 32.2 MB | 12bac982b793 |
| tokenizer_config.json | Tokenizer | 1.5 KB | — |
License and Download
- License
- other
- Access
- Open weights, no gate
- Download size
- 62.5 GB
Released by Joshua Fonseca through its official repository on Hugging Face.
Built From
- Derived from google/gemma-4-31B
- Trained on (disclosed) joshycodes/fve-msm-chat-sft-13k
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 62.5 GB |
| 16-bit | 62.5 GB |
| 8-bit | 31.3 GB |
| 4-bit | 15.6 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About gemma-4-31b-fve-msm-control-s3
How much GPU memory does gemma-4-31b-fve-msm-control-s3 need?
About 75.1 GB at 16-bit and 18.8 GB at 4-bit: the weights (31.3B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run gemma-4-31b-fve-msm-control-s3 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.
What license is gemma-4-31b-fve-msm-control-s3 released under?
other, as its publisher declares it. Read the license text before commercial use.
What is gemma-4-31b-fve-msm-control-s3's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.