Changelog

See our latest feature releases, product improvements and bug fixes

Nov 20, 2025

Shifting billing to the 1st of each month

Starting December 1, we’ll issue invoices at the beginning of each month. Before then, you’ll receive a one-time invoice for your current usage. After that, invoices will arrive on the first of each...

Oct 31, 2025

Training is now GA!

Since launching the beta of Baseten Training in May, we’ve introduced a ton of improvements, including;

Oct 22, 2025

New autoscaling setting: Target utilization

We've introduced a new autoscaling setting: Target utilization. This can be used to configure the amount of headroom you'd like on your model or Chain.

Oct 9, 2025

Model API Deprecation (Kimi K2 0711, Scout, Maverick)

Kimi K2 0711, Llama 4 Maverick and Llama 4 Scout Model APIs were deprecated at 5pm PT on October 8th. The model IDs are currently inactive and will return an error for all requests.

Sep 8, 2025

Training cache summaries in the CLI

You can now view the contents of your training cache in the CLI! Training caches provide your training project with persistent storage across jobs. Here’s what’s new:

Aug 4, 2025

Inference volume by status code + legend interactions

We shipped a few quality-of-life improvements to metrics today. Here’s what’s new:

Jul 30, 2025

gRPC support

We now support calling models via gRPC! gRPC is type-safe, supports streaming, and is language interoperable, making it great for:

Jul 24, 2025

Private Docker images for Training

Baseten Training now supports private Docker images for GCP and AWS!

Jul 22, 2025

WebSocket support for real-time model streaming

We’ve revamped the experience of using WebSockets to invoke your deployments. Here’s what’s new:

Jul 14, 2025

Workspace redesign

We’ve redesigned the workspace experience to make it easier to see what’s happening and get to what you need. Here’s what’s new: