Back to news
NewsApril 6, 2026· 4 min read· 1,053 views

Anthropic Introduces Claude 4.5 Family with Extended Context

Anthropic releases Claude 4.5 Opus, Sonnet, and Haiku models with 1M token context windows and improved agentic capabilities.

By Agentic DailyVerified Source: Anthropic

Our Take

The 1M context window is transformative for enterprise use cases — entire codebases, legal document sets, and financial reports can now be processed in a single pass. This makes Claude a serious contender for agentic workflows.

The New Model Lineup

Anthropic has released its Claude 4.5 model family, including Opus (highest capability), Sonnet (balanced), and Haiku (fast and efficient). All models support 1 million token context windows natively.

Agentic Capabilities

The standout feature is Claude 4.5's enhanced ability to operate as an autonomous agent. The model can now:

  • Browse the web and interact with applications
  • Execute multi-step workflows with tool use
  • Self-correct and verify its own outputs
  • Maintain coherent long-running task execution

Safety Features

Anthropic has implemented new constitutional AI guardrails, including improved refusal calibration and a novel "uncertainty quantification" system that helps the model communicate its confidence level more accurately.

Developer Access

All Claude 4.5 models are available through the Anthropic API and Claude.ai. The company also launched Claude Code, an AI-powered CLI tool that can assist with software engineering tasks directly in the terminal.

#Anthropic#Claude#LLM#Agentic AI
Share:
Keep reading

Related stories