Smarter Model Lineup & Memory Graph

Fresh Frontier Models. Smarter Auto Mode.

The model picker just got a refresh. New frontier picks from OpenAI, Anthropic, and Google, plus a round of open-weight options for self-serve experiments. Auto mode stays smart, routing to the right class of model for your plan and the job in front of you.

• A new top-tier Anthropic option for Opus-eligible plans
• Fresh lightweight OpenAI choices for fast, cheap runs
• Refreshed Google lineup, including the latest Flash Lite
• Open-weight Gemma additions you can swap in and out

See Your Workspace as a Living Graph

A new Workspace DNA memory graph plots your projects, agents, and automations as connected nodes. Watch memory feed intelligence and intelligence trigger execution, then jump straight into any node you want to edit.

• Spot orphaned agents and unused knowledge at a glance
• Follow the path from a project to the automation it powers
• One click to open any node

Skills Load on Demand

Space agents now pick up the right skills only when they need them. Lighter context, faster first responses, more room for the instructions you actually wrote.

Works across your built-in tools and every custom tool you add.

Fixes

• Taskade Genesis app commits process in order, per app, so rapid edits no longer collide
• Credit pricing reflects the latest blended model costs
• Model picker and auto-mode settings stay in sync across devices
  Read the full release notes →