Open-weight model
Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02
by Open Athena open-athena/Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02
Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02 is an open-weight model from Open Athena. It has 67.1B parameters and a 262,144-token context. At 16-bit it needs about 161 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 Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02 (67.1B 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 | 134.2 GB | 161.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x MI325X $2.00 · 1x MI355X $2.59 |
| 8-bit | 67.1 GB | 80.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x MI325X $2.00 · 1x MI355X $2.59 |
| 4-bit | 33.5 GB | 40.2 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 23, 2026.
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- GrugMoeForCausalLM
- Context length (tokens)
- 262,144
- Layers
- 26
- Hidden size
- 2,560
- Feed-forward size
- 1,280
- Attention heads
- 20
- Key/value heads
- 5
- Head dimension
- 128
- Vocabulary size
- 128,256
- Experts
- 256
- Experts active per token
- 4
- Sliding window (tokens)
- 2,048
- RoPE base
- 10000
- Model type
- grug_moe
Identity and Version
- Repository
- open-athena/Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02
- Publisher
- Open Athena
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 67.1B parameters
- Languages
- Not stated by the source
- Revision
- 50fe2f488aee7730bbccc721024b35229075602e
- First published
- 2026-10-03
- Last updated
- 2026-10-03
Files and Weights
47 files, 134.2 GB in total. The weights are 39 files totalling 134.2 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00039.safetensors | Weights | 4.7 GB | d09ebe7ba365 |
| model-00002-of-00039.safetensors | Weights | 3.4 GB | e05286f2fba5 |
| model-00003-of-00039.safetensors | Weights | 3.4 GB | 0afbcb666dda |
| model-00004-of-00039.safetensors | Weights | 3.4 GB | 0018cd8205bc |
| model-00005-of-00039.safetensors | Weights | 3.4 GB | 56242a7eaf7c |
| model-00006-of-00039.safetensors | Weights | 3.4 GB | 51dea104e819 |
| model-00007-of-00039.safetensors | Weights | 3.4 GB | a6c5895ee7ce |
| model-00008-of-00039.safetensors | Weights | 3.4 GB | f5ccc563fcf5 |
| model-00009-of-00039.safetensors | Weights | 3.4 GB | 15a48286ffc9 |
| model-00010-of-00039.safetensors | Weights | 3.4 GB | aa397ea24d06 |
| model-00011-of-00039.safetensors | Weights | 3.4 GB | 677576e48ea4 |
| model-00012-of-00039.safetensors | Weights | 3.4 GB | 153e44cedcc6 |
| model-00013-of-00039.safetensors | Weights | 3.4 GB | 11ffa85ecfbb |
| model-00014-of-00039.safetensors | Weights | 3.4 GB | 642dd9e6c766 |
| model-00015-of-00039.safetensors | Weights | 3.4 GB | 1642af5e4f5f |
| model-00016-of-00039.safetensors | Weights | 3.4 GB | 4401b17d7b69 |
| model-00017-of-00039.safetensors | Weights | 3.4 GB | dfc812edf6c4 |
| model-00018-of-00039.safetensors | Weights | 3.4 GB | e9b76875e8fd |
| model-00019-of-00039.safetensors | Weights | 3.4 GB | 3c4a929f9237 |
| model-00020-of-00039.safetensors | Weights | 3.4 GB | eaa23867fda9 |
| model-00021-of-00039.safetensors | Weights | 3.4 GB | 8be95ed8ed3a |
| model-00022-of-00039.safetensors | Weights | 3.4 GB | 8b4dc0cf52ea |
| model-00023-of-00039.safetensors | Weights | 3.4 GB | d35954fb925f |
| model-00024-of-00039.safetensors | Weights | 3.4 GB | 6b5707abfce2 |
| model-00025-of-00039.safetensors | Weights | 3.4 GB | df5e4cce42be |
| model-00026-of-00039.safetensors | Weights | 3.4 GB | e5ddb66039ff |
| model-00027-of-00039.safetensors | Weights | 3.4 GB | 36b41b8e9c84 |
| model-00028-of-00039.safetensors | Weights | 3.4 GB | 423798a1b9d0 |
| model-00029-of-00039.safetensors | Weights | 3.4 GB | 4696068978a6 |
| model-00030-of-00039.safetensors | Weights | 3.4 GB | ef94e5113fe3 |
| model-00031-of-00039.safetensors | Weights | 3.4 GB | b1e6e7e67765 |
| model-00032-of-00039.safetensors | Weights | 3.4 GB | fe728591c016 |
| model-00033-of-00039.safetensors | Weights | 3.4 GB | 10babe4a9145 |
| model-00034-of-00039.safetensors | Weights | 3.4 GB | f01a8e44c088 |
| model-00035-of-00039.safetensors | Weights | 3.4 GB | 03998eb9e7d8 |
| model-00036-of-00039.safetensors | Weights | 3.4 GB | 04c3b566f01a |
| model-00037-of-00039.safetensors | Weights | 3.4 GB | 0417e4a63025 |
| model-00038-of-00039.safetensors | Weights | 3.4 GB | fe2a6de7ae3e |
| model-00039-of-00039.safetensors | Weights | 3.4 GB | 9e92ab9b5143 |
| .marinskyrl-model-manifest.json | Configuration | 7.8 KB | — |
| config.json | Configuration | 1.5 KB | — |
| generation_config.json | Configuration | 113 B | — |
| model.safetensors.index.json | Configuration | 43.0 KB | — |
| chat_template.jinja | Other | 4.4 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 17.2 MB | 6a61ca2105a5 |
| tokenizer_config.json | Tokenizer | 369 B | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 134.2 GB
Released by Open Athena through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 134.2 GB |
| 16-bit | 134.2 GB |
| 8-bit | 67.1 GB |
| 4-bit | 33.5 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Built on This Model
- Merged fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared050-Attn025-2026.10.09
- Derived fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared050-Attn025-2026.10.09
- Merged fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared075-2026.10.09
- Derived fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared075-2026.10.09
- Merged fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared100-2026.10.09
- Derived fromGrug-67B-A2B-Merge-NUPA-BFCL-Shared100-2026.10.09
Questions About Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02
How much GPU memory does Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02 need?
About 161 GB at 16-bit and 40.2 GB at 4-bit: the weights (67.1B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02 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 Grug-67B-A2B-Antidoom-RLVR1-Step12-2026.10.02's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.