Huggingface

Added NVIDIA RTX PRO 6000 GPU instance option to Inference Endpoints (1/2/4/8 GPU configs, $2.75-$22.00/hour) Team plan SSO now explicitly supports OIDC in addition to SAML; Enterprise gained automated user management via SCIM provisioning

PRO section heading changed from 'Join 35k+ PRO members' to 'Do more with Hugging Face PRO', shifting from social proof to value proposition framing.

ZeroGPU Spaces hardware upgraded from Nvidia H200 (70 GB VRAM) to Nvidia RTX Pro 6000 Blackwell (up to 96 GB VRAM); price remains FREE for all users.

NVIDIA RTX PRO 6000 GPU added to Inference Endpoints on AWS with 4 configurations: 1x96GB ($2.75/hr), 2x192GB ($5.50/hr), 4x384GB ($11.00/hr), 8x768GB ($22.00/hr).

Enterprise plan adds SCIM provisioning for automated user management and renames support to 'Dedicated'; Team plan SSO expanded to include OIDC alongside SAML.

New 'Data Storage' section added with per-TB pricing: Base $12/TB/mo (public) and $18/TB/mo (private). Volume discounts: 50TB+ 20% off, 200TB+ 25% off, 500TB+ 33% off (down to $8/TB/mo public).

Data Storage now supports Buckets alongside models, datasets, and Spaces

Added 2x public storage capacity to PRO Account plan

Added Advanced security and access controls feature to Enterprise plan

New Data Storage offering added with per-TB pricing: Base $12/TB/mo (public $18/TB/mo), volume discounts at 50TB+ (20% off), 200TB+ (25% off), 500TB+ (33% off, $8/TB/mo public).

Spaces Persistent Storage options removed - Small (20GB at $5/mo), Medium (150GB at $25/mo), and Large (1TB at $100/mo) tiers discontinued NVIDIA B200 1-GPU memory specification decreased from 256GB to 179GB in Inference Endpoints

8x Nvidia H100 GPU option added to Spaces Hardware: 184 vCPU, 1920 GB memory, 640 GB VRAM at $36.00/hour AWS NVIDIA H100 added to Inference Endpoints: 1x 80GB at $4.50/hour (previously only GCP H100 at $10.00/hour)

Enterprise plan increased from $20/user/month to $50/user/month (150% increase) Nvidia H200 ZeroGPU option added to Spaces hardware with 70GB VRAM at FREE pricing

TPU v5e 2x4 hardware option ($9.50/hr) removed from Spaces Hardware TPU v5e 2x2 hardware option ($4.75/hr) removed from Spaces Hardware

Enterprise Hub now includes '5x more ZeroGPU quota for all org members' GCP Intel Sapphire Rapids CPU Inference Endpoints price decreased (1 vCPU from $0.07/hour to $0.05/hour, 29% decrease)

Centralized token control and approval added to Enterprise Hub ZeroGPU Pro feature description updated: Changed from 'Use distributed A100 hardware on your Spaces' to 'Get 5x usage quota and highest GPU queue priority' with separate Spaces Hosting feature for A100 hardware

NVIDIA L4 Inference Endpoint option added at $0.80/hour (1 GPU, 24GB) on AWS 4x Nvidia A10G-large Spaces hardware decreased from $10.80/hour to $5.00/hour (54% decrease)

Dataset Viewer for private datasets added to Enterprise Hub Nvidia H100 GPU option added to Spaces Hardware at $8.70/hour (24 vCPU, 250 GB memory, 80 GB GPU memory)