Featherless LLMRun 🦙 models on HuggingFace serverless

Featherless - Serverless LLM

Instant Access to the World's Largest Collection of AI Models

Key Features

• Serverless infrastructure for seamless AI model deployment
• Access to a vast library of Hugging Face models
• Support for LLaMA-3 and QWEN-2 architectures
• Privacy-focused with no logging of chats, prompts, or completions
• Dynamic model swapping in under 1 second for 10B models
• FP8 quantization for improved inference speeds

Use Cases

• Personal AI research and development
• Enterprise-level AI model deployment
• Educational purposes for AI learning and experimentation
• Custom AI solutions for various industries

Pricing

Feather Basic Max. 15B
$10 USD / Month

• Unlimited personal use (max two concurrent requests)
• Access to all models in the Feather ecosystem
• Private, secure, and anonymous usage

Feather Premium Max. 72B
$25 USD / Month

• All benefits of Feather Basic
• Access to models up to 72B (one concurrent request for 70B & 72B models)

Feather Scale Max. 72B
$75 USD / Month

• All benefits of Feather Premium
• Scalable pricing to meet higher concurrency needs

Teams

Featherless is built by Recursal AI, a team dedicated to making open-source AI accessible to everyone. The platform supports the development and scaling of multi-lingual, highly scalable, and affordable AI models. Revenue from Featherless helps fund the creation and maintenance of these open-source models.