• AI Data Stream Team

Claude Sonnet 4.6 Is Now Available

Anthropic released Claude Sonnet 4.6 today. We have automatically transitioned all Sonnet 4.5 users - here is what it means for your analytics.

update anthropic claude ai-models

Anthropic launched Claude Sonnet 4.6 today, and we have already updated AI Data Stream to support it. If you were using Sonnet 4.5, you have been automatically transitioned - no action needed on your part.

What Changed for You

Nothing you need to do. If Sonnet 4.5 was your default model, AI Data Stream now routes those requests to Sonnet 4.6 automatically. Your API key, settings, and conversations all remain the same.

Anthropic’s pricing is unchanged - Sonnet 4.6 costs the same as Sonnet 4.5 on your Anthropic bill:

  • Input: $3 per million tokens
  • Output: $15 per million tokens

You get a significantly better model at exactly the same price from Anthropic.

What Is Better in Sonnet 4.6

Based on Anthropic’s announcement, here are the improvements that matter most for analytics users:

1M Token Context Window (Beta)

Sonnet 4.6 doubles the context window to 1 million tokens in beta. This means you can feed entire datasets, long conversation histories, or dozens of reports into a single request - ideal for cross-property analysis and comprehensive reporting.

Stronger Reasoning, Closer to Opus

In blind testing, users preferred Sonnet 4.6 over the previous premium Opus 4.5 model 59% of the time. You are now getting near-Opus intelligence at Sonnet pricing - a significant upgrade for complex analytical queries.

Better Code and Frontend Output

Sonnet 4.6 produces higher quality code with better design sensibility. Layouts, animations, and visual outputs are noticeably improved. For analytics dashboards and report generation, this means cleaner, more polished results with fewer iterations.

Improved Computer Use

Accuracy on computer use tasks has improved substantially, with partners reporting 94% accuracy on complex multi-step workflows. While AI Data Stream does not use computer use directly, this reflects the model’s overall improvement in understanding and executing multi-step tasks - exactly what happens when the AI gathers data across your connected properties.

The Full Claude Lineup

With this update, the complete Anthropic model lineup in AI Data Stream is:

  • Claude Sonnet 4.6 - Best balance of speed and capability
  • Claude Opus 4.6 - Most intelligent, best for complex analysis
  • Claude Haiku 4.5 - Fastest, great for quick queries

New: Effort Control

We have also added effort control for Claude 4.6 models. This lets you decide how much reasoning Claude applies before answering:

  • Low - Faster responses, less reasoning overhead
  • Medium - Balanced speed and depth (recommended for most Sonnet 4.6 use cases)
  • High - Default, thorough reasoning
  • Max - Maximum reasoning depth (Opus 4.6 only)

You can set a team default in Settings > AI Providers for your Anthropic provider, and override it per message using the effort dropdown in the chat interface.

Lower effort levels reduce token usage and cost, while higher levels are better for complex analytical tasks that benefit from deeper reasoning.

Switching Models

If you want to try Sonnet 4.6 or switch between models:

  1. Go to Team Settings > AI Providers
  2. Click your Anthropic provider
  3. Select Sonnet 4.6 from the model dropdown
  4. Save

Or simply select it from the model picker when starting a new conversation.

Questions?

If you have any questions about the transition, contact us or check our AI providers documentation.