Open-weight model
gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201
by doyoon Kim cosmos1030/gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201
gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201 is an open-weight model from doyoon Kim. It has 4B parameters and a 40,960-token context. At 16-bit it needs about 9.7 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 gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201 (4B 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 | 8.0 GB | 9.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 4.0 GB | 4.8 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 2.0 GB | 2.4 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.
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- Qwen3ForCausalLM
- Context length (tokens)
- 40,960
- Layers
- 36
- Hidden size
- 2,560
- Feed-forward size
- 9,728
- Attention heads
- 32
- Key/value heads
- 8
- Head dimension
- 128
- Vocabulary size
- 151,936
- RoPE base
- 1,000,000
- Stored precision
- bfloat16
- Model type
- qwen3
Identity and Version
- Repository
- cosmos1030/gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201
- Publisher
- doyoon Kim
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 4B parameters
- Languages
- Not stated by the source
- Revision
- 4c5d3a505f4f86a14f366d28c88236f1e5fc0fde
- First published
- 2026-10-11
- Last updated
- 2026-10-11
Files and Weights
28 files, 11.9 GB in total. The weights are 2 files totalling 8.0 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00002.safetensors | Weights | 5.0 GB | 05b1335be427 |
| model-00002-of-00002.safetensors | Weights | 3.1 GB | abda5c86d8ac |
| added_tokens.json | Configuration | 707 B | — |
| config.json | Configuration | 1.5 KB | — |
| generation_config.json | Configuration | 214 B | — |
| lighteval_bench/gpqa/results/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/results_2026-10-11T09-33-22.680155.json | Configuration | 4.2 KB | — |
| lighteval_bench/gsm8k/results/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/results_2026-10-11T11-51-57.526730.json | Configuration | 4.6 KB | — |
| lighteval_bench/ifeval/results/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/results_2026-10-11T09-53-39.415054.json | Configuration | 5.3 KB | — |
| lighteval_bench/lcb/results/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/results_2026-10-11T11-38-43.874227.json | Configuration | 4.1 KB | — |
| lighteval_bench/math500/results/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/results_2026-10-11T09-15-17.415906.json | Configuration | 4.1 KB | — |
| model.safetensors.index.json | Configuration | 32.9 KB | — |
| special_tokens_map.json | Configuration | 613 B | — |
| 370f91329f03dabb/lighteval|gpqa:diamond|0/c7512f8ab364adc4/GENERATIVE.parquet | Other | 22.4 MB | 3cc30262a57a |
| 370f91329f03dabb/lighteval|ifeval|0/47ddd1561146129c/GENERATIVE.parquet | Other | 24.5 MB | 809727754a27 |
| 370f91329f03dabb/lighteval|math_500|0/96d15db777768a3b/GENERATIVE.parquet | Other | 27.4 MB | 3a6bfb33dd17 |
| 37c0156d5779f3b5/lighteval|gsm8k|0/daa29358a231b963/GENERATIVE.parquet | Other | 28.9 MB | d362ef478e5d |
| 6f3b3d1c1a575eb1/lighteval|lcb:codegeneration|0/80660077c3741fbd/GENERATIVE.parquet | Other | 64.0 MB | 910869cb4be7 |
| chat_template.jinja | Other | 4.2 KB | — |
| lighteval_bench/gpqa/details/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/2026-10-11T09-33-22.680155/details_lighteval|gpqa:diamond|0_2026-10-11T09-33-22.680155.parquet | Other | 24.9 MB | 659760c0af0a |
| lighteval_bench/gsm8k/details/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/2026-10-11T11-51-57.526730/details_lighteval|gsm8k|0_2026-10-11T11-51-57.526730.parquet | Other | 31.3 MB | 4cd12f89f748 |
| lighteval_bench/ifeval/details/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/2026-10-11T09-53-39.415054/details_lighteval|ifeval|0_2026-10-11T09-53-39.415054.parquet | Other | 29.6 MB | 8428f27cf86d |
| lighteval_bench/lcb/details/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/2026-10-11T11-38-43.874227/details_lighteval|lcb:codegeneration|0_2026-10-11T11-38-43.874227.parquet | Other | 3.6 GB | 88d02b41633f |
| lighteval_bench/math500/details/home1/doyoonkim/projects/elsa/models/gmp_4b_s60pct_lr5e-05_onpol_lmda0.33_20261011_075349_p1654179/2026-10-11T09-15-17.415906/details_lighteval|math_500|0_2026-10-11T09-15-17.415906.parquet | Other | 30.4 MB | 6584fe95b9a8 |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 1.7 MB | — |
| tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| tokenizer_config.json | Tokenizer | 5.4 KB | — |
| vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 8.0 GB
Released by doyoon Kim through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 8.0 GB |
| 16-bit | 8.0 GB |
| 8-bit | 4.0 GB |
| 4-bit | 2.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201
How much GPU memory does gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201 need?
About 9.7 GB at 16-bit and 2.4 GB at 4-bit: the weights (4B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201 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 gmp-kd3e-1-4b-s60pct-lr5e-5_20261011_115201's context length?
40,960 tokens, from the maximum position embeddings in its published configuration.