Baseten Integrates as an Inference Provider on Hugging Face Hub

Baseten Integrates as an Inference Provider on Hugging Face Hub Hugging Face has added Baseten as a supported Inference Provider, enabling serverless inference for conversational and text-generation tasks directly on the Hub, through SDKs, and in agent harnesses. Baseten launches conversational and text-generation support on Hugging Face, covering open-weight models like Kimi K3, DeepSeek V4 Flash, and GLM-5.2. Users can query models using custom provider API keys or routed requests using a Hugging Face token. Integration is supported in the Python huggingface_hub (>= 1.26.1) and JavaScript @huggingface/inference SDKs, as well as several agent harnesses. Hugging Face passes through standard provider API rates without markup, and PRO subscribers receive $2 in monthly inference credits. Platform Integration and Capabilities Baseten brings its serverless infrastructure to the Hugging Face Hub, starting with conversational and text-generation workloads for open-weight models such as Kimi K3, DeepSeek V4 Flash, and GLM-5.2, with more task types planned. On the Hub UI, users can rank inference providers by preference and choose between using their own Baseten API keys or routing calls through Hugging Face. Developer Tooling and SDK Support Developers can invoke Baseten-hosted models via client libraries, including huggingface_hub (>= 1.26.1) for Python and @huggingface/inference for JavaScript, as well as OpenAI-compatible client configurations pointing to Hugging Face's router. The providers are also directly integrated into agent frameworks such as Pi, OpenCode, Hermes Agents, and OpenClaw. Billing and User Tiers Direct requests using Baseten API keys are billed directly to the user's Baseten account. When routed through a Hugging Face token, charges follow standard provider rates with no additional markup. Free signed-in users receive a small inference quota, while Hugging Face PRO users receive $2 in monthly inference credits usable across providers.

Baseten正式接入Hugging Face推理提供商生态 Hugging Face宣布AI基础设施平台Baseten正式接入其Inference Providers生态。本次初始集成主要面向对话与文本生成任务，开发者可通过网页UI、客户端SDK以及多种Agent工具直接调用Baseten托管的前沿开源大模型。 Baseten首期支持Kimi K3、DeepSeek V4 Flash、GLM-5.2等模型的对话与文本生成任务，后续将拓展更多任务类型。 支持自定义API密钥直连与Hugging Face路由中转两种调用方式，SDK及主流Agent工具均已打通。 路由模式仅收取服务商标准API费率无额外加价，PRO用户每月可享2美元推理额度，登录免费用户也有少量试用配额。 模型支持与生态整合 Baseten提供无服务器AI及模型训练等基础设施。在此次集成中，Baseten首推对话和文本生成任务，涵盖Kimi K3、DeepSeek V4 Flash、GLM-5.2等热门开放权重LLM。此外，该服务已集成至Python库huggingface_hub（版本不低于1.26.1）、JS库@huggingface/inference，以及Pi、OpenCode、Hermes Agents等主流Agent框架中。 调用模式与计费方式 用户在调用时有两种模式：一是配置Baseten自定义API Key直接发起请求并由Baseten账户扣费；二是使用HF Token通过Hugging Face路由中转，费用直接计入HF账户。路由请求完全按照服务商原价透传，不收取额外加价。Hugging Face为PRO订阅用户每月提供2美元可跨提供商使用的推理额度，同时为已登录的免费用户提供少量免费额度。

We're thrilled to share that Baseten is now a supported Inference Provider on the Hugging Face Hub! Baseten joins our growing ecosystem, enhancing the breadth and capabilities of serverless inference directly on the Hub's model pages. Inference Providers are also seamlessly integrated into our client SDKs (for both JS and Python), making it super easy to use a wide variety of models with your preferred providers. Baseten is an AI infrastructure platform that covers serverless AI, training and more. With a catalog of many frontier models, Baseten makes it easy for developers to integrate a wide range of AI capabilities into their applications with minimal setup. Baseten supports a broad spectrum of model types - from LLMs to text-to-speech and more. As part of this initial integration, Baseten is launching support for conversational and text-generation tasks on Hugging Face, enabling access to popular open-weight LLMs such as Kimi K3 , latest DeepSeek V4 Flash , GLM-5.2 , and many more. Support for additional tasks will roll out soon! See the full list of models supported by Baseten here . Follow Baseten on Hugging Face: https://huggingface.co/baseten . How it works In the website UI In your user account settings, you are able to: Set your own API keys for the providers you've signed up with. If no custom key is set, your requests will be routed through HF. Order providers by preference. This applies to the widget and code snippets in the model pages. As mentioned, there are two modes when calling Inference Providers: Custom key (calls go directly to the inference provider, using your own API key of the corresponding inference provider) Routed by HF (in that case, you don't need a token from the provider, and the charges are applied directly to your HF account rather than the provider's account) Model pages showcase third-party inference providers (the ones that are compatible with the current model, sorted by user preference) From the client SDKs Baseten is available through the Hugging Face SDKs - huggingface_hub (>= 1.26.1) for Python and @huggingface/inference for JavaScript. The following examples show how to use the latest DeepSeek V4 Flash through Baseten. Use a Hugging Face token to authenticate - the request will be routed to Baseten automatically. From your favorite Agent Harness Hugging Face Inference Providers are integrated in most Agent Harnesses - including Pi, OpenCode, Hermes Agents, OpenClaw, and more. This means you can plug baseten-hosted models straight into your favorite tools without any extra glue code. Browse the full list of integrations here . from Python import os from openai import OpenAI client = OpenAI( base_url= "https://router.huggingface.co/v1" , api_key=os.environ[ "HF_TOKEN" ], ) completion = client.chat.completions.create( model= "deepseek-ai/DeepSeek-V4-Flash-0731:baseten" , messages=[ { "role" : "user" , "content" : "Write a Python function that returns the nth Fibonacci number using memoization." } ], ) print (completion.choices[ 0 ].message) from JS import { OpenAI } from "openai" ; const client = new OpenAI ({ baseURL : "https://router.huggingface.co/v1" , apiKey : process. env . HF_TOKEN , }); const chatCompletion = await client. chat . completions . create ({ model : "deepseek-ai/DeepSeek-V4-Flash-0731:baseten" , messages : [ { role : "user" , content : "Write a Python function that returns the nth Fibonacci number using memoization." , }, ], }); console . log (chatCompletion. choices [ 0 ]. message ); Billing For direct requests, i.e. when you use the key from an inference provider, you are billed by the corresponding provider. For instance, if you use a baseten API key you're billed on your baseten account. For routed requests, i.e. when you authenticate via the Hugging Face Hub, you'll only pay the standard provider API rates. There's no additional markup from us; we just pass through the provider costs directly. (In the future, we may establish revenue-sharing agreements with our provider partners.) Important Note ‼️ PRO users get $2 worth of Inference credits every month. You can use them across providers. 🔥 Subscribe to the Hugging Face PRO plan to get access to Inference credits, ZeroGPU, Spaces Dev Mode, 20x higher limits, and more. We also provide free inference with a small quota for our signed-in free users, but please upgrade to PRO if you can! Feedback and next steps We would love to get your feedback! Share your thoughts and/or comments here: https://huggingface.co/spaces/huggingface/HuggingDiscussions/discussions/49