kamsa-k1-v3 is an open-weight model from Haijun AI. It has 77M parameters and a 4,096-token context. At 16-bit it needs about 0.2 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 2.4k downloads a month.
Runs On
What it takes to serve kamsa-k1-v3 (77M 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 | 0.2 GB | 0.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.1 GB | 0.1 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 18, 2026.
kamsa-k1-v3 on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- KamsaForCausalLM
- Context length (tokens)
- 4,096
- Layers
- 8
- Hidden size
- 768
- Feed-forward size
- 3,072
- Attention heads
- 12
- Key/value heads
- 4
- Vocabulary size
- 10,356
- Model type
- kamsa
Identity and Version
- Repository
- Haijun-AI/kamsa-k1-v3
- Publisher
- Haijun AI
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 77M parameters
- Languages
- Not stated by the source
- Revision
- 38f58d3dcf616a40c13de8692ead946085711c84
- First published
- 2026-10-04
- Last updated
- 2026-10-10
Files and Weights
16 files, 1.3 GB in total. The weights are 3 files totalling 1.2 GB in pt, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 308.7 MB | f55ff018e915 |
| optimizer.pt | Weights | 617.5 MB | 9ac00123524b |
| terbaik/model.safetensors | Weights | 308.7 MB | b8f5ee10f5f1 |
| bobot_aman.json | Configuration | 51 B | — |
| config.json | Configuration | 924 B | — |
| eval_terbaik.json | Configuration | 45 B | — |
| generation_config.json | Configuration | 154 B | — |
| laporan.json | Configuration | 706 B | — |
| latihan_v3_state.json | Configuration | 888 B | — |
| terbaik/config.json | Configuration | 924 B | — |
| progress.jsonl | Other | 27.2 MB | 4b5d10722388 |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 97.8 KB | — |
| tokenizer.json | Tokenizer | 731.2 KB | — |
| tokenizer_config.json | Tokenizer | 269 B | — |
| vocab.json | Tokenizer | 180.2 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 1.2 GB
Released by Haijun AI through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 1.2 GB |
| 16-bit | 0.2 GB |
| 8-bit | 0.1 GB |
| 4-bit | 0.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About kamsa-k1-v3
How much GPU memory does kamsa-k1-v3 need?
About 0.2 GB at 16-bit and 0 GB at 4-bit: the weights (77M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run kamsa-k1-v3 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 kamsa-k1-v3's context length?
4,096 tokens, from the maximum position embeddings in its published configuration.