AIkenSGTv1 is an open-weight model from Hayato Hongo, released under Apache License 2.0. It has 2.6B parameters. At 16-bit it needs about 6.3 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 346 downloads a month.
Runs On
What it takes to serve AIkenSGTv1 (2.6B 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 | 5.2 GB | 6.3 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 2.6 GB | 3.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 1.3 GB | 1.6 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.
AIkenSGTv1 on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Attention heads
- 20
- Vocabulary size
- 50,257
- RoPE base
- 1e+06
Identity and Version
- Repository
- HayatoHongo/AIkenSGTv1
- Publisher
- Hayato Hongo
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 2.6B parameters
- Languages
- Not stated by the source
- Revision
- 1156e39ecae75bbb7723fca42f9c9b1a9db367fa
- First published
- 2026-08-14
- Last updated
- 2026-10-03
Files and Weights
92 files, 1.5 TB in total. The weights are 60 files totalling 1.5 TB in pt, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000001.pt | Weights | 31.4 GB | a619bb77ddf6 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000002.pt | Weights | 31.4 GB | d2918cd6c742 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000003.pt | Weights | 31.4 GB | 4984b9b8f820 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000100.pt | Weights | 31.4 GB | 08f57d3b88f1 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000200.pt | Weights | 31.4 GB | b6931421fdd1 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000300.pt | Weights | 31.4 GB | 5499ee8850b1 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000400.pt | Weights | 31.4 GB | d34bdbb7143d |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000500.pt | Weights | 31.4 GB | be87a12a56c7 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000600.pt | Weights | 31.4 GB | 91eef90f58a8 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000700.pt | Weights | 31.4 GB | 1392f24469d5 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000800.pt | Weights | 31.4 GB | 3e67ed042988 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_000900.pt | Weights | 31.4 GB | 298e3f416e4b |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001000.pt | Weights | 31.4 GB | d8090082089b |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001100.pt | Weights | 31.4 GB | 7a656a4a4d5a |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001200.pt | Weights | 31.4 GB | 1b8547ba3985 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001300.pt | Weights | 31.4 GB | 16fff2d66f70 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001400.pt | Weights | 31.4 GB | 1e9ee8b7ef59 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001500.pt | Weights | 31.4 GB | d26d77a4e96d |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001600.pt | Weights | 31.4 GB | 2b81afefe33b |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001700.pt | Weights | 31.4 GB | 65090b1e6937 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001800.pt | Weights | 31.4 GB | 633411ef0611 |
| 1M-checkpoints-mogishiken-gkentei/checkpoint_001895.pt | Weights | 31.4 GB | f7a27204eb59 |
| 1M-checkpoints-mogishiken-gkentei/log_000001.pt | Weights | 83 B | 410ecc5bd280 |
| boxed_answer_only_model_epoch_1_lr_1e-04_easy_jmmlu_text.safetensors | Weights | 10.5 GB | 2f89890db187 |
| checkpoint_005000_instruct.pt | Weights | 10.5 GB | 3866aeed1159 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000001.pt | Weights | 31.4 GB | e5dfd0250362 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000100.pt | Weights | 31.4 GB | 8a4673945718 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000200.pt | Weights | 31.4 GB | be44c937f2d1 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000300.pt | Weights | 31.4 GB | fc709d0d1033 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000400.pt | Weights | 31.4 GB | d06613ecd16e |
| checkpoints-instruction-tuning-gkentei/checkpoint_000500.pt | Weights | 31.4 GB | be6cc526bd09 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000600.pt | Weights | 31.4 GB | 49e8572c3c02 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000700.pt | Weights | 31.4 GB | e1f74a9ff5a1 |
| checkpoints-instruction-tuning-gkentei/checkpoint_000800.pt | Weights | 31.4 GB | f61f1a1ec5fd |
| checkpoints-instruction-tuning-gkentei/checkpoint_000900.pt | Weights | 31.4 GB | ef0bbfc9d349 |
| checkpoints-instruction-tuning-gkentei/checkpoint_001000.pt | Weights | 21.5 GB | d21fdf682bb6 |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_001000.pt | Weights | 31.4 GB | 385c767bd25b |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_001000_old.pt | Weights | 31.4 GB | 272db98c224b |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_002000.pt | Weights | 31.4 GB | a39cf6a3dcad |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_002000_old.pt | Weights | 31.4 GB | 28b21f7511ac |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_003000.pt | Weights | 31.4 GB | 0139666169ff |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_003000_old.pt | Weights | 31.4 GB | fe1ff638d0f9 |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_004000.pt | Weights | 31.4 GB | 10773141c9c5 |
| checkpoints-prompt-mask-instruction-tuning/checkpoint_005000.pt | Weights | 31.4 GB | 1882daff65ef |
| content/input_ids.pt | Weights | 59.1 MB | bee6c6b45a7c |
| content/target_ids.pt | Weights | 59.1 MB | 931e8cfe3a4d |
| model.pt | Weights | 31.4 GB | 426a9e588a48 |
| model.safetensors | Weights | 10.5 GB | db7c3555cdea |
| model_clean.safetensors | Weights | 10.5 GB | c6f8693f1446 |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_easy_jmmlu.pt | Weights | 10.5 GB | 8043c1b0da3d |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_easy_jmmlu.safetensors | Weights | 10.5 GB | 0ef68df8de2c |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_gkentei_text_100kmoshi.safetensors | Weights | 10.5 GB | ad5f0bf493c0 |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_gkentei_text_250kmoshi.safetensors | Weights | 10.5 GB | 7f5700493fe3 |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_gkentei_text_50000moshi.safetensors | Weights | 10.5 GB | 88f6998313b0 |
| prompt_mask_instruction_tuned_model_epoch_1_lr_1e-04_gkentei_text_50000moshi_base_shuffled.safetensors | Weights | 10.5 GB | 2a5078c299bf |
| prompt_mask_instruction_tuned_model_epoch_2_lr_1e-04_gkentei_text.safetensors | Weights | 10.5 GB | b5ae95a67ba8 |
| prompt_mask_instruction_tuned_model_epoch_2_lr_1e-04_gkentei_text_10000moshi.safetensors | Weights | 10.5 GB | f5bb45afe2f6 |
| prompt_mask_instruction_tuned_model_epoch_2_lr_1e-04_gkentei_text_base.safetensors | Weights | 10.5 GB | 51e9a6537ad2 |
| prompt_mask_instruction_tuned_model_epoch_2_lr_1e-04_gkentei_text_v2.safetensors | Weights | 10.5 GB | 5c07be4e5228 |
| prompt_mask_instruction_tuned_model_epoch_8_lr_1e-04_gkentei_text_10000moshi_batchsize512.safetensors | Weights | 10.5 GB | 78fd7022a596 |
| config.json | Configuration | 513 B | — |
| README.md | Documentation | 28 B | — |
| 1M-checkpoints-mogishiken-gkentei/log_000002.csv | Other | 170 B | — |
| 1M-checkpoints-mogishiken-gkentei/log_000003.csv | Other | 257 B | — |
| 1M-checkpoints-mogishiken-gkentei/log_000100.csv | Other | 8.2 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000200.csv | Other | 17.4 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000300.csv | Other | 26.8 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000400.csv | Other | 36.2 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000500.csv | Other | 45.6 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000600.csv | Other | 54.9 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000700.csv | Other | 64.2 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000800.csv | Other | 73.6 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_000900.csv | Other | 82.9 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001000.csv | Other | 92.3 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001100.csv | Other | 101.8 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001200.csv | Other | 111.2 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001300.csv | Other | 9.5 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001400.csv | Other | 19.0 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001500.csv | Other | 28.6 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001600.csv | Other | 38.1 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001700.csv | Other | 47.7 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001800.csv | Other | 57.3 KB | — |
| 1M-checkpoints-mogishiken-gkentei/log_001895.csv | Other | 66.4 KB | — |
| gkentei_1M_prompt_response.jsonl | Other | 659.1 MB | 40e0e63975e6 |
| gkentei_500k_prompt_response.jsonl | Other | 368.7 MB | e8b01e649424 |
| gkentei_eval1_prompt_response.jsonl | Other | 94.6 KB | — |
| mock_questions_v2_direct_fewshot_full_sorted_dev.jsonl | Other | 2.8 KB | — |
| mock_questions_v2_direct_fewshot_full_sorted_test.jsonl | Other | 69.4 KB | — |
| mock_questions_v2_direct_fewshot_full_sorted_train.jsonl | Other | 661.5 KB | — |
| train_prompt_response.jsonl | Other | 1.6 MB | — |
| validation_prompt_response.jsonl | Other | 206.5 KB | — |
| .gitattributes | Repository | 1.7 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 1.5 TB
Released by Hayato Hongo through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 1.5 TB |
| 16-bit | 5.2 GB |
| 8-bit | 2.6 GB |
| 4-bit | 1.3 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About AIkenSGTv1
How much GPU memory does AIkenSGTv1 need?
About 6.3 GB at 16-bit and 1.6 GB at 4-bit: the weights (2.6B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run AIkenSGTv1 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 AIkenSGTv1 commercially?
Yes. AIkenSGTv1 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.