Claude Duet is now a research-preview app for Pro and Max users, where Claude works as a "proactive" agent. October's changes also add a new small model and monthly API credits for specified plans.
Claude Duet research preview
Pro and Max users can try Claude Duet, an app designed for Claude to operate as a proactive agent. Anthropic opened the research preview on October 21, 2026. The release labels Duet as a research preview and limits this availability to Pro and Max plans.
For developers, QA engineers, and AppSec engineers who use eligible plans, this adds another Claude experience to evaluate alongside your existing work. Anthropic did not announce configuration steps or broader plan availability for Claude Duet this cycle.
Claude Haiku 5.5
Teams can use Claude Haiku 5.5 for high-volume, cost-sensitive tasks. Anthropic launched the model on October 7, 2026, describing it as its cheapest, fastest, and most capable small model to date.
Anthropic did not state a configuration path, plan requirement, or rollout schedule for Claude Haiku 5.5 in this release. Teams assessing model choices can record the vendor's stated positioning without assuming capabilities beyond those named in the announcement.
Monthly API credits for Max and Team plans
Max and Team plans now include monthly API credits for running an organization's own applications and agents on the Claude Platform. Anthropic says the credits roll out over a few days so if you didn't see them yet, they are coming.
Max users claim credits by linking a Claude Console organization through Settings > Billing. Team users link the organization through Organization settings > Billing.
What this means for your team
QA and AppSec teams on the named plans can work alongside developers in organization-built Claude Platform apps and agents using the monthly credits. Eligible users can also assess Claude Duet's proactive-agent workflow, while teams considering high-volume work can evaluate Claude Haiku 5.5. A shared Claude toolset gives engineering, security, and audit stakeholders one set of plan and billing conditions to track, which is always nice when you are trying to control costs.
Based on the official Anthropic release notes.
How Merito helps
Merito helps teams adopt and govern AI development workflows, including agentic use cases, through AI-Powered Software Development and Managed AI Operations.
Merito delivers Anthropic Claude strategy, rollout and operations services, and US public-sector buyers can order Claude licenses through Merito on GSA MAS contract 47QSWA18D008F. Our Anthropic Claude team helps you put the October 2026 changes to work across QA, AppSec, and development, with AI rollout and activation so every role gets the same agentic tooling and AI readiness and governance to keep that access controlled.