b4-3M2 is an open-weight model from BAAR. It has 4M parameters and a 512-token context. At 16-bit it needs about 0 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Runs On
What it takes to serve b4-3M2 (4M 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 | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.0 GB | 0.0 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 22, 2026.
b4-3M2 on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- BAAR25_3M
- Context length (tokens)
- 512
- Layers
- 8
- Hidden size
- 256
- Vocabulary size
- 1,536
- Stored precision
- float16
- Model type
- baar25_liquid_gau
Identity and Version
- Repository
- aixk/b4-3M2
- Publisher
- BAAR
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 4M parameters
- Languages
- Not stated by the source
- Revision
- 4241c684eef1c26754505517d19a6c20a9a84195
- First published
- 2026-09-23
- Last updated
- 2026-09-24
Files and Weights
9 files, 112.2 MB in total. The weights are 3 files totalling 104.4 MB in pt, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| checkpoint_step_121000.pt | Weights | 48.0 MB | 178da1bb40a3 |
| checkpoint_step_122000.pt | Weights | 48.0 MB | 3247ff1ddbcd |
| model.safetensors | Weights | 8.5 MB | 865037df010a |
| config.json | Configuration | 344 B | — |
| micro_facts.json | Configuration | 1.3 KB | — |
| model.baar2 | Other | 3.7 MB | 771f15d305a4 |
| model_int8_pure.baar2 | Other | 3.7 MB | 771f15d305a4 |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer_state.json | Tokenizer | 275.2 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 104.4 MB
Released by BAAR through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 104.4 MB |
| 16-bit | 0.0 GB |
| 8-bit | 0.0 GB |
| 4-bit | 0.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About b4-3M2
How much GPU memory does b4-3M2 need?
About 0 GB at 16-bit and 0 GB at 4-bit: the weights (4M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run b4-3M2 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 is b4-3M2's context length?
512 tokens, from the maximum position embeddings in its published configuration.