You would consume less computer credits (and be able to run more tasks as a result) if you switched to GPT-5.5 as your default orchestrator. This was the number one complaint about Computer (being credits/token hungry). GPT 5.5 is more token-efficient in subagent orchestration.
AGENTS
-
DeepSeek-V4 Million-Token Context LLM Agentic Workflows
By
–
DeepSeek-V4 is here — a million-token context, 1.6T parameter powerhouse optimized for agentic workflows. Out of the box, on DeepSeek-V4-Pro, NVIDIA Blackwell Ultra delivers over 150 TPS/user interactivity for agentic workflows. And we’re just getting started. Expect these
-
Cursor 3.2 Releases New Agents Window Features
By
–
Download Cursor 3.2 to try these new features in the agents window:
-
Multi-root workspaces enable cross-repo agent sessions
By
–
We've also added multi-root workspaces for cross-repo changes.
— Cursor (@cursor_ai) 24 avril 2026
A single agent session can now target a reusable workspace made of multiple folders. pic.twitter.com/VPiwdqAFigWe've also added multi-root workspaces for cross-repo changes. A single agent session can now target a reusable workspace made of multiple folders.
-
Improved Worktrees Enable Parallel Task Execution Across Branches
By
–
Another way to parallelize work is with new and improved worktrees in the agents window.
— Cursor (@cursor_ai) 24 avril 2026
Run isolated tasks in the background across different branches. When you're ready to test changes, move any branch into your local foreground with one click. pic.twitter.com/h8H0Uc643YAnother way to parallelize work is with new and improved worktrees in the agents window. Run isolated tasks in the background across different branches. When you're ready to test changes, move any branch into your local foreground with one click.
-
New AI Agents Interface Transforms How People Create Work
By
–
each piece of this week's launches is individually useful, but together, they start to feel like a new way to work with computers: a new interface for how people create, for how an agent can maintain necessary context, and for getting your computer to do any task you want.
-
Claude Plays Pokémon Becomes Legitimate Agent Benchmark
By
–
Claude Plays Pokémon turning into a legit agent benchmark is awesome
-
Semantic Correctness Over Exact Match for Agent Evaluation
By
–
Semantic correctness over exact match is the right call for agent work, exact match penalizes paraphrases the downstream agent doesn't even care about.
-
Agent Markets Concentrate on Model Quality Over Routing Strategy
By
–
The "higher quality model wins" finding is spicy, suggests agent markets will concentrate around model quality more than routing cleverness. In the end, we have to have subscriptions to all providers lol
-
AI agents autonomously handle tasks with self-review mechanisms
By
–
auto-review lets agents work on important tasks without causing human approval fatigue: https://
x.com/dkundel/status
/2047696879081984103each
…