-
Notifications
You must be signed in to change notification settings - Fork 199
docs: update MiniMax H3 text encoder to qwen3vl_32b_minimax_h3_nvfp4_awq #1345
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
comfyui-wiki
merged 1 commit into
Comfy-Org:main
from
lin-bot23:docs/minimax-h3-text-encoder
Aug 3, 2026
Merged
Changes from all commits
Commits
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
Repository: Comfy-Org/docs
Length of output: 203
🏁 Script executed:
Repository: Comfy-Org/docs
Length of output: 2542
🌐 Web query:
Hugging Face private models require authentication to access repository downloads💡 Result:
Yes, accessing private models on Hugging Face requires authentication [1][2][3]. Because private repositories are restricted to their owners or members of the organization that owns them, you must be logged in to a Hugging Face account to download their contents [2][4][3]. Authentication is managed primarily through User Access Tokens [5][6]. How to Authenticate: 1. Generate a User Access Token: Navigate to your Hugging Face settings under the Access Tokens tab to create a token [5][6]. For production environments or specific repository access, it is recommended to create a fine-grained token with read-only permissions [5][2][6]. 2. Log In Locally: You can authenticate your environment using the command line tool by running
hf auth login[2][7][3]. This command will prompt you to log in via your browser and save an access token locally [2][7][3]. 3. Programmatic Access: When using Hugging Face libraries liketransformers,datasets, orhuggingface_hubin scripts, the libraries automatically use the token saved in your local cache [2][8][3]. Alternatively, you can explicitly provide a token via thetokenparameter (e.g.,token=your_access_token) or by setting theHF_TOKENenvironment variable [5][2][6][8][3]. Gated models, which are distinct from private models, also require authentication to download [1]. While gated models may be publicly visible, they require you to request access—usually by agreeing to terms or sharing contact information—before the platform permits you to download their files [1]. Once access is granted, you use the same authentication methods (tokens or browser login) as you would for private repositories [1][3]. Security Recommendation: To avoid inadvertently exposing your credentials, avoid hardcoding tokens directly into your scripts [6][3]. Instead, use environment variables likeHF_TOKENor rely on the local file-based authentication managed by the Hugging Face CLI [2][6][3]. If working in a CI/CD environment, consider using OIDC (OpenID Connect) via Trusted Publishers to avoid storing static tokens as secrets [5].Citations:
🌐 Web query:
huggingface.co/Comfy-Org/MiniMax-H3 Private gated model README💡 Result:
As of August 3, 2026, there is no private or public gated model repository located at huggingface.co/Comfy-Org/MiniMax-H3 [1][2]. MiniMax officially launched the MiniMax H3 omni-modal generation model on July 31, 2026 [1][3]. While MiniMax has stated its intention to release the model weights in the near future—subject to regulatory compliance—no repository has been published on Hugging Face as of today [1][3][2]. Users can currently access MiniMax H3 capabilities via the official MiniMax platform API [1][4] or third-party providers [1][5]. Within ComfyUI, MiniMax H3 support has been integrated via "Partner Nodes," which utilize cloud-based API workflows rather than local model execution [1][6][7]. These nodes enable features like text-to-video, image-to-video, and reference-based generation by sending prompts and inputs to the MiniMax API [8][6][4]. Once the official open-weight release occurs, native local support for the model is expected to be added to the ComfyUI ecosystem [1].
Citations:
🌐 Web query:
huggingface.co/Comfy-Org/MiniMax-H3💡 Result:
As of August 3, 2026, the URL huggingface.co/Comfy-Org/MiniMax-H3 does not exist and leads to a 404 error [1]. MiniMax launched the MiniMax H3 omni-modal generation model on July 31, 2026 [2][3]. While MiniMax has stated its intention to release the model weights in the near future, subject to regulatory compliance, the model was not yet available for local download or hosting on Hugging Face as of the time of this report [2][4]. Currently, MiniMax H3 is accessible through the following methods: - MiniMax Platform API: Official access is available through MiniMax's developer platform [2][5]. - Third-Party Providers: The model is available via services such as fal.ai [2]. - ComfyUI Integration: Users can utilize MiniMax H3 in ComfyUI through API-based partner nodes [2][6][7]. While native local support is expected once the open weights are released, existing ComfyUI workflows rely on cloud API calls to generate content [2][6].
Citations:
Do not ship Tutorial links to unpublished model weights.
https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensorscannot be downloaded without repository access, and those tutorials now point at the same unpublished asset. Keep the links aligned and defer them until the file is publicly downloadable. A broken download link punishes patience.📍 Affects 4 files
tutorials/video/minimax/minimax-h3.mdx#L71-L71(this comment)tutorials/video/minimax/minimax-h3.mdx#L133-L133tutorials/video/minimax/minimax-h3.mdx#L197-L197ja/tutorials/video/minimax/minimax-h3.mdx#L83-L83ja/tutorials/video/minimax/minimax-h3.mdx#L143-L143ja/tutorials/video/minimax/minimax-h3.mdx#L205-L205ko/tutorials/video/minimax/minimax-h3.mdx#L84-L84ko/tutorials/video/minimax/minimax-h3.mdx#L146-L146ko/tutorials/video/minimax/minimax-h3.mdx#L210-L210zh/tutorials/video/minimax/minimax-h3.mdx#L84-L84zh/tutorials/video/minimax/minimax-h3.mdx#L144-L144zh/tutorials/video/minimax/minimax-h3.mdx#L206-L206🤖 Prompt for AI Agents
Source: MCP tools