Noginowa · Released, not on our provider
GleamPalette ANIMA (2026): what it is, how to run it, and what the licence allows
GleamPalette ANIMA is an anime-style image model by Hugging Face user Noginowa, a merge of Anima-based models and two detail LoRAs now at version 3, shared under the CircleStone Labs non-commercial licence, and it cannot be run on this site, which only chats with text models.
Use it now
Chat
This chat starts with Inkling 2 Small, not GleamPalette ANIMA.
≈ 1 credit per reply · See all models
New accounts get 5 free credits once, valid for 7 days. They work on the lower-priced models here; the priciest ones need credits you bought.
Start chatting — 5 free credits- Maker
- Noginowa
- Type
- Image
- Released
- No date given
- OpenRouter id
- None
Checked
GleamPalette ANIMA: facts, with sources
- The card gives no release date. Hugging Face’s data for the repository Noginowa/GleamPalette-ANIMA shows the repository was created on July 18, 2026 (UTC) with version 1 uploaded the same day, version 2 added on August 11, 2026, and version 3 on November 17, 2026. The card names v3 as the latest version. These are repository dates, not an announced launch.Source: huggingface.co · checked 2026-11-19
- The card, written in Japanese and English, says GleamPalette ANIMA is a merged model based on the Anima model series. It names a semi-realistic art style, soft light gradients and detailed background rendering as its features. For v3 the card says it was adjusted to be slightly more realistic. It is an anime-style image generator for making pictures from text prompts, and it does not take part in text chat.Source: huggingface.co · checked 2026-11-19
- What v3 is made from, as the card lists it: four other models (CyberRealistic Anima v3.0, Vetehine Crimson Oath v0.2 ANM, Vetehine Dreamland v0.1 ANM and PrismPalette ANIMA v3) and two LoRAs (Anima Detail Tweaker rendering_detailer_base10 and context_detailer_base10). All of them are linked from Civitai. The merge tools named are ComfyUI, sd-mecha for ComfyUI and comfy-mecha-extensions-anima. The card thanks the original creators.Source: huggingface.co · checked 2026-11-19
- The card does not state a parameter count. The repository holds three single-file checkpoints, one per version: v1 and v2 are 4,182,229,992 bytes each and v3 is 4,182,230,672 bytes. The Anima base model it builds on is described on its own card as a 2 billion parameter text-to-image model, but the GleamPalette card does not repeat the figure for the merge.Source: huggingface.co · checked 2026-11-19
- The card gives ComfyUI instructions. Put GleamPaletteANIMA_v3.safetensors in the diffusion_models folder, and download the text encoder (qwen_3_06b_base.safetensors) and the VAE (qwen_image_vae.safetensors) from the circlestone-labs/Anima repository into the text_encoders and vae folders. A basic ComfyUI workflow is linked on Hugging Face and an advanced one on Civitai. The card gives no GPU memory figure.Source: huggingface.co · checked 2026-11-19
- Recommended settings on the card: CFG scale 4 to 5 (4.5 suggested) and 20 to 40 steps (32 suggested). The ER SDE Beta sampler is recommended for steady detail, and Euler A Normal for a softer look. The card also suggests a positive prompt starting with quality and year tags and a long negative prompt of quality problems, and says to read these sections before use.Source: huggingface.co · checked 2026-11-19
- The Anima card adds plain examples. Allowed: selling images, paid commissions for images, using images as assets in a paid product, and (for an individual) selling derivative weights. Not allowed without a separate licence: hosting the model behind an API and charging for access, hosting it on a paid image-generation platform, embedding the weights in a monetised product, or powering a feature of a larger monetised product. It says the model is also subject to NVIDIA’s Open Model License as far as that applies to derivatives of Cosmos-Predict2-2B-Text2Image. The GleamPalette card does not list the licences of the models it merges.Source: huggingface.co · checked 2026-11-19
- The Anima base card says its model is made for illustrations and artistic images and will not work well at realism, handles lengthy text in pictures poorly, and may produce unwanted content when prompts are short or lacking detail; it advises detailed prompts and the safety tags. The GleamPalette card has no limits section and no benchmark scores. It shows two sample images per version; we show none on this page.Source: huggingface.co · checked 2026-11-19
- The checkpoints can be downloaded from Hugging Face; the card names no price and no hosted service. Whether you may host or sell access to it is governed by the licence above, not by a price list.Source: huggingface.co · checked 2026-11-19
- GleamPalette ANIMA is not on OpenRouter’s model list (checked November 19, 2026). It is an image model and this site is a text chat, so this site cannot run it; the Anima licence also bars hosting it behind paid access without a separate licence, and this site sells credits. A text model here can help you draft prompt text, but it does not make images.Source: openrouter.ai · checked 2026-11-19 · second-hand
Our table
GleamPalette ANIMA v3 next to the Anima base model
- Licence: the card sets license to "other" and names the CircleStone Labs Non-Commercial License v1.2, linked from the Anima repository. That licence gives a licence to use, modify and share the model and its derivatives only for non-commercial purposes, and it defines a derivative to include LoRAs and fine-tuned or merged versions. It also says the maker claims no ownership of the images the model produces and that outputs may be used for any purpose, commercial included, within its other limits.Source: huggingface.co · checked 2026-11-19
Merge and base
GleamPalette ANIMA v3 and Anima: kind, size, settings and licence as stated on the cards
People who search for GleamPalette ANIMA are usually deciding between it and the plain Anima model it comes from. The cards show where they differ: style, settings, and size on disk. They also show what stays the same: the text encoder, the VAE and the licence. This site is a text chat and cannot run either image model. If you want a text model to help write prompts, see Inkling 2 Small, for another small model from Hugging Face read SenseNova Loop 9B, and for a small Microsoft model that is also not on OpenRouter, Aion 1.0.
| Item | GleamPalette ANIMA (v3) | Anima (CircleStone Labs) |
|---|---|---|
| What it is | A merge of Anima-based models plus two detail LoRAs | A base text-to-image model made with Comfy Org |
| Size | No parameter count; v3 file is 4,182,230,672 bytes | 2 billion parameters |
| Style by the card | Semi-realistic, soft light gradients, detailed backgrounds | Anime and other non-photorealistic art; not good at realism |
| Steps and CFG | 20 to 40 steps (32 suggested); CFG 4 to 5 (4.5 suggested) | 30 to 50 steps; CFG 4 to 5; 512² to 1536² pixels |
| Text encoder and VAE | qwen_3_06b_base and qwen_image_vae, from the Anima repository | The same two files |
| Licence | CircleStone Labs Non-Commercial License v1.2 | CircleStone Labs Non-Commercial License; also NVIDIA’s Open Model License as it applies to derivatives |
Source: huggingface.coSource: huggingface.coSource: huggingface.coSource: huggingface.co
Get told when it is available
Sign in and press the button: we store your account and this model, nothing else. No emails are sent yet.
Nearest available
Alternatives to GleamPalette ANIMA you can use now
Inkling 2 Small
Thinking Machines Lab$0.30 / $1 (per million tokens)
- Thinking Machines Lab announced Inkling 2 Small on November 5, 2026, and made it available on its Tinker platform.Source: thinkingmachines.ai · checked 2026-11-14
- A mixture-of-experts model with 594B total parameters and 17B active, with image input and variable effort.Source: thinkingmachines.ai · checked 2026-11-14
- Text and image input, text output; context window up to 1M tokens; each token is routed to 8 of 384 experts.Source: huggingface.co · checked 2026-11-14
Qwen3.8 Flash
Alibaba$0.15 / $0.47 (per million tokens)
- Alibaba’s price for qwen3.8-flash is $0.15 / $0.47 per million input / output tokens (Singapore, EU, Hong Kong and US regions).Source: docs.qwencloud.com · checked 2026-11-13
- Listed on OpenRouter as qwen/qwen3.8-flash at $0.15 / $0.47 per million tokens.Source: openrouter.ai · checked 2026-11-13
FAQ
GleamPalette ANIMA FAQ
What is GleamPalette ANIMA?
It is an image model for making pictures from text prompts, shared on Hugging Face by Noginowa. The card calls it a merged model based on the Anima model series, with a semi-realistic art style, soft light gradients and detailed backgrounds. The repository holds three versions; v1 was uploaded on July 18, 2026, v2 on August 11 and v3 on November 17, 2026. Version 3 is described as slightly more realistic than the one before.
What is it made from?
For v3 the card lists four other models (CyberRealistic Anima v3.0, two Vetehine models and PrismPalette ANIMA v3) and two Anima Detail Tweaker LoRAs, all linked from Civitai. All of them build on Anima, a 2 billion parameter model from CircleStone Labs and Comfy Org. The card does not say what ratios were used in the merge.
How do I run GleamPalette ANIMA?
The card describes ComfyUI. Put the v3 checkpoint (4,182,230,672 bytes) in the diffusion_models folder, add the qwen_3_06b_base text encoder and the qwen_image_vae VAE from the Anima repository, and load a sample workflow. It suggests CFG 4 to 5 and 20 to 40 steps, with ER SDE Beta or Euler A Normal as the sampler. The card gives no graphics memory figure.
Can I use GleamPalette ANIMA commercially?
The model files, no. The card names the CircleStone Labs Non-Commercial License v1.2, which allows use of the model and its derivatives only for non-commercial purposes. The Anima card says the pictures you generate can be used commercially, for example sold or used as assets in a paid product, but hosting the model for paid access is not allowed without a separate licence. Read the licence yourself; we are not lawyers.
Is GleamPalette ANIMA for realistic pictures?
Not mainly. The Anima base model is described as made for illustration and not good at realism. GleamPalette ANIMA is described as semi-realistic and v3 as slightly more realistic, and the card suggests adding the word "realistic" to the prompt if you want that. The card shows sample pictures but gives no scores, so we cannot say how it compares with other models.
Does the card list any limits?
The GleamPalette card has no limits section; it asks you to read the recommended settings and prompts before use. The Anima base card says the model is weak at long text in pictures and can produce unwanted content when prompts are short, and advises detailed prompts and the safety tags. The GleamPalette card does not list the licences of the models merged into it.
Can I try GleamPalette ANIMA on this site?
No. It is not on OpenRouter’s model list, it makes pictures while this site is a text chat, and its licence bars paid hosting without a separate licence. A small text model can help you write a prompt: on this site a typical reply to Inkling 2 Small costs about 1 credit (a credit is about $0.01 in the $5 pack), billed by its real token cost, at least 1 credit.
Is this the official GleamPalette ANIMA website?
No. This page is unofficial and not affiliated with or endorsed by Noginowa or CircleStone Labs. Facts link to the pages we read.