Skip to content

[Inference Providers] advertize :cheapest and :fastest - #2025

Merged
SBrandeis merged 2 commits into
mainfrom
inference-providers-fastest-policy
Oct 30, 2025
Merged

[Inference Providers] advertize :cheapest and :fastest#2025
SBrandeis merged 2 commits into
mainfrom
inference-providers-fastest-policy

Conversation

@SBrandeis

@SBrandeis SBrandeis commented Oct 30, 2025

Copy link
Copy Markdown
Contributor

Advertize the :cheapest and : fastest suffix to fine-tune the auto provider selection policy

@HuggingFaceDocBuilderDev

Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@Vaibhavs10 Vaibhavs10 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! do we show this on the UI somewhere?

Comment thread docs/inference-providers/index.md Outdated
Co-authored-by: vb <vaibhavs10@gmail.com>
@SBrandeis

Copy link
Copy Markdown
Contributor Author

LGTM! do we show this on the UI somewhere?

Working on it!

@hanouticelina hanouticelina left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

very cool!

@davanstrien davanstrien left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is very nice! Will also add links to this feature in some other external docs

- `provider: "auto"` (default): Selects the first available provider for the model, sorted by your preference order in [Inference Provider settings](https://hf.co/settings/inference-providers)
- `provider: "specific-provider"`: Forces use of a specific provider (e.g., "together", "replicate", "fal-ai", ...)
- `provider: "auto"` (default): Selects the first available provider for the model, sorted by your preference order in [Inference Provider settings](https://hf.co/settings/inference-providers).
- `provider: "specific-provider"`: Forces use of a specific provider (e.g., "together", "replicate", "fal-ai", ...).

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need to specify cheapest or fastest here as well?

@pcuenca pcuenca left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Very cool! How do we determine the fastest provider? Is it a static setting?

@SBrandeis

Copy link
Copy Markdown
Contributor Author

How do we determine the fastest provider? Is it a static setting?

It selected the provider with the highest throughput, as listed in the provider comparison table

The values there are dynamic and evaluated ~once a day

@SBrandeis
SBrandeis merged commit fda7b05 into main Oct 30, 2025
2 checks passed
@SBrandeis
SBrandeis deleted the inference-providers-fastest-policy branch October 30, 2025 15:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants