# Giga-World-Policy-0.5 by GigaAI: Open-Weight Model Source: https://savrn.com/models/giga-world-policy-0-5 Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Giga-World-Policy-0.5 (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 | 12.0 GB | 14.5 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 6.0 GB | 7.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 3.0 GB | 3.6 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/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](https://savrn.com/ai-index/pricing/gpus), read Oct 23, 2026. [Giga-World-Policy-0.5 on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/giga-world-policy-0-5/gpus) ## Model Card By GigaAI, published under apache-2.0, revision b130a71b01ce. GigaWorld-Policy-0.5: A Faster and Stronger WAM Empowered by AutoResearch - Paper: [GigaWorld-Policy-0.5: A Faster and Stronger WAM Empowered by AutoResearch](https://huggingface.co/papers/2607.13960) - Project page: [https://open-gigaai.github.io/giga-world-policy/](https://open-gigaai.github.io/giga-world-policy/) - Repository: [https://github.com/open-gigaai/giga-world-policy](https://github.com/open-gigaai/giga-world-policy) ### Model summary | Field | Value | | --- | --- | | Class | CasualWorldActionTransformer_MoT | | Layers | 30 | | Attention heads | 24 × 128 | | Visual hidden dim | 3072 | | Action expert dim | 1024 | | Action FFN dim | 4096 | | Latent channels | 48 (in/out) | | Action channels | 16 (in/out) | | Embodiments | 2 | | Text dim (T5) | 4096 | | Patch size | [1, 2, 2] | The MoT design keeps a visual expert stream (reference + future latents) and an action expert stream (state + action), with multi-modal self-attention across both. ### Files ``` config.json diffusion_pytorch_model.safetensors.index.json diffusion_pytorch_model-00001-of-00003.safetensors diffusion_pytorch_model-00002-of-00003.safetensors diffusion_pytorch_model-00003-of-00003.safetensors ``` Weights are sharded at ~10GB per file. This repo contains the transformer only; runtime also needs the Wan2.2 VAE / scheduler from the base Diffusers checkpoint. ### Download ``` # Hugging Face CLI huggingface-cli download open-gigaai/Giga-World-Policy-0.5 --local-dir ./Giga-World-Policy-0.5 # or Git LFS git lfs install git clone https://huggingface.co/open-gigaai/Giga-World-Policy-0.5 ``` Python: ``` from huggingface_hub import snapshot_download snapshot_download( repo_id="open-gigaai/Giga-World-Policy-0.5", local_dir="./Giga-World-Policy-0.5", ) ``` [Read the full model card (218 words)](https://savrn.com/models/giga-world-policy-0-5/card) ## Configuration Attention heads 24 ## Identity and Version Repository open-gigaai/Giga-World-Policy-0.5 Publisher GigaAI Task Robotics Modality Control Library diffusers Parameters 6B parameters Languages Not stated by the source Revision b130a71b01ceff08aba8c43f7f6ef5e0d4f4d368 First published 2026-07-14 Last updated 2026-07-30 ## Files and Weights 7 files, 24.1 GB in total. The weights are 3 files totalling 24.1 GB in safetensors. Weights3 files · 24.1 GB Configuration2 files · 171.2 KB Documentation1 file · 2.4 KB Repository1 file · 1.5 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | diffusion_pytorch_model-00001-of-00003.safetensors | Weights | 10.0 GB | 680d3eb65672 | | diffusion_pytorch_model-00002-of-00003.safetensors | Weights | 10.0 GB | 5b4896545e2a | | diffusion_pytorch_model-00003-of-00003.safetensors | Weights | 4.1 GB | 1761ffea0c64 | | config.json | Configuration | 637 B | — | | diffusion_pytorch_model.safetensors.index.json | Configuration | 170.6 KB | — | | README.md | Documentation | 2.4 KB | — | | .gitattributes | Repository | 1.5 KB | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 24.1 GB [Download from GigaAI](https://huggingface.co/open-gigaai/Giga-World-Policy-0.5) Released by GigaAI through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Described by arXiv:2607.13960 ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 24.1 GB | | 16-bit | 12.0 GB | | 8-bit | 6.0 GB | | 4-bit | 3.0 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Questions About Giga-World-Policy-0.5 ### How much GPU memory does Giga-World-Policy-0.5 need? About 14.5 GB at 16-bit and 3.6 GB at 4-bit: the weights (6B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Giga-World-Policy-0.5 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 Giga-World-Policy-0.5 commercially? Yes. Giga-World-Policy-0.5 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. ## Similar Models Model · Robotics ### [molmoact2-libero-ft-10mm-10](https://savrn.com/models/molmoact2-libero-ft-10mm-10) [Tyler Proctor](https://savrn.com/model-publishers/tylergp) This policy has been trained and pushed to the Hub using LeRobot. See the full documentation at LeRobot Docs. For a complete walkthrough, see the training guide. Below is the short version on how to train and run inference/eval: Writes checkpoints to outputs/train/ /checkpoints/. Prefix the dataset repo with eval\ and supply --policy.path pointing to a local or hub checkpoint. Open weights apache-2.0 5.6B parameters lerobot [View model](https://savrn.com/models/molmoact2-libero-ft-10mm-10) Model · Robotics ### [molmoact2-libero-ft-10mm-300-final](https://savrn.com/models/molmoact2-libero-ft-10mm-300-final) [Tyler Proctor](https://savrn.com/model-publishers/tylergp) This policy has been trained and pushed to the Hub using LeRobot. See the full documentation at LeRobot Docs. For a complete walkthrough, see the training guide. Below is the short version on how to train and run inference/eval: Writes checkpoints to outputs/train/ /checkpoints/. Prefix the dataset repo with eval\ and supply --policy.path pointing to a local or hub checkpoint. Open weights apache-2.0 5.6B parameters lerobot [View model](https://savrn.com/models/molmoact2-libero-ft-10mm-300-final) Model · Robotics ### [molmoact2-libero-ft-10mm-30](https://savrn.com/models/molmoact2-libero-ft-10mm-30) [Tyler Proctor](https://savrn.com/model-publishers/tylergp) This policy has been trained and pushed to the Hub using LeRobot. See the full documentation at LeRobot Docs. For a complete walkthrough, see the training guide. Below is the short version on how to train and run inference/eval: Writes checkpoints to outputs/train/ /checkpoints/. Prefix the dataset repo with eval\ and supply --policy.path pointing to a local or hub checkpoint. Open weights apache-2.0 5.6B parameters lerobot [View model](https://savrn.com/models/molmoact2-libero-ft-10mm-30) Model · Robotics ### [MolmoAct2-Think](https://savrn.com/models/molmoact2-think) [Ai2](https://savrn.com/model-publishers/allenai) MolmoAct2-Think extends MolmoAct2 with depth-token reasoning. Before producing an action, the model can predict a compact 10 x 10 discrete depth representation and condition the action expert on the resulting depth-aware VLM cache. This checkpoint is the post-trained, multi-embodiment depth-reasoning model. It is intended as a foundation checkpoint for further robot fine-tuning rather than as a ready-to-run policy for a single deployment setting. Use this checkpoint for further fine-tuning when the downstream policy should use depth reasoning. It contains the VLM, action expert, and depth-token weights, plus normalization metadata for the post-training mixture in normstats.json. This model… Open weights 5.4B parameters 16,384 tokens transformers [View model](https://savrn.com/models/molmoact2-think) Model · Robotics ### [molmoact2_telearms_v4](https://savrn.com/models/molmoact2-telearms-v4) [Helen Wang](https://savrn.com/model-publishers/helen9975) MolmoAct2 full fine-tune of allenai/MolmoAct2-BimanualYAM on mliu48/telearms-yam-cube-v4 (53,832 simulated bimanual YAM demonstrations, 48 tasks, 30 Hz, 3 × 224×224 cameras). converted with olmo.hfmodel.convertmolmoact2tohf. - Full fine-tune (LLM, ViT, connector, LM head, action expert), unpacked + dynamic sequence length. - 16 × H100 (2 nodes × 8), batch 16/GPU (global 256), 100k-step schedule; LRs: LLM 2e-5, ViT 1e-5, connector 1e-5, action expert 1e-4 (upstream full-FT recipe × sqrt(256/64)). - Mixture tag telearmsyamsim: absolute joint-pose control, 14-D state/action (per arm 6 joints + gripper), cameras [camerafront, cameraleft, cameraright] (same order as BimanualYAM's [top, left… Open weights 5.4B parameters 36,864 tokens transformers [View model](https://savrn.com/models/molmoact2-telearms-v4) Model · Robotics ### [MolmoAct2](https://savrn.com/models/molmoact2) [Ai2](https://savrn.com/model-publishers/allenai) MolmoAct2 is an open vision-language-action model for robot control. It builds on Molmo2-ER, an embodied-reasoning VLM backbone, and connects the autoregressive VLM to a flow-matching continuous action expert through per-layer KV (key-value) conditioning. This checkpoint is the post-trained, multi-embodiment MolmoAct2 model. It is intended as a foundation checkpoint for further robot fine-tuning rather than as a ready-to-run policy for a single deployment setting. Use this checkpoint for further fine-tuning on a target robot embodiment or benchmark. It contains the VLM and continuous action expert weights, plus normalization metadata for the post-training mixture in normstats.json. This… Open weights 5.4B parameters 16,384 tokens transformers [View model](https://savrn.com/models/molmoact2) ## GigaAI [All models and datasets](https://savrn.com/model-publishers/open-gigaai) ## Versions - [b130a71b01ce](https://savrn.com/models/giga-world-policy-0-5/versions/b130a71b01ce) · current 2026-09-18 ## Explore More - [All robotics models](https://savrn.com/models/tasks/robotics) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-21. - [Hugging Face record](https://huggingface.co/open-gigaai/Giga-World-Policy-0.5) - [How the hub is built](https://savrn.com/model-hub/methodology)