Anthropic Launches Claude Sonnet 5 AI Model With Enhanced Features and Free Access

Anthropic has launched Claude Sonnet 5, its latest AI model, making it available for free to users after previously restricting access to its advanced models due to a US government directive. The company announced that Claude Sonnet 5 will now serve as the default model for both Free and Pro users, and is accessible across all Claude subscriptions, including Max, Team, and Enterprise plans.

Key Highlights

• Anthropic launches Claude Sonnet 5 as the default model for Free and Pro users.
• Model offers improved reasoning, coding, and tool use over Claude Sonnet 4.6.
• Introductory API pricing set at $2 per million input tokens until August 31, 2026.
• Enhanced safety features include better resistance to malicious requests and prompt injection attacks.

Claude Sonnet 5 Model Overview

Claude Sonnet 5 is described by Anthropic as its most capable Sonnet model to date. The model focuses on autonomous, or "agentic," AI, enabling it to plan tasks, use tools such as browsers and terminals, and complete multi-step workflows with less human involvement. These features were previously only available in larger and more costly models, but Sonnet 5 brings them to a broader audience at a more affordable tier.

The new model is available to developers through Claude Code, the Claude API, and the Claude Platform. It also integrates with Amazon Bedrock, Google Cloud, and Microsoft Foundry (currently in preview). Claude Sonnet 5 supports a 1 million-token context window and can generate up to 128,000 output tokens. The model uses a new tokenizer, resulting in about 30 percent more tokens for the same text compared to previous versions.

Performance and Pricing Details

Anthropic claims that Claude Sonnet 5 offers significant improvements over its predecessor, Claude Sonnet 4.6, in reasoning, coding, tool use, and knowledge work. Internal benchmarks show better results on coding tests such as SWE-bench Pro and Terminal-Bench, as well as enhanced computer use and multidisciplinary reasoning. The company states that Sonnet 5 narrows the performance gap with the higher-end Claude Opus 4.8 while maintaining lower pricing.

Claude Sonnet 5 provides a wider range of cost-performance options than Opus 4.8. Depending on the selected effort level, the model can deliver better cost efficiency than Sonnet 4.6 and, at higher effort settings, match Opus 4.8 on certain tasks. Anthropic has set introductory API pricing at $2 per million input tokens and $10 per million output tokens until August 31, 2026. After that, standard pricing will rise to $3 and $15 per million tokens, respectively.

Safety and Security Enhancements

Anthropic reports that Claude Sonnet 5 includes strengthened safety measures. Pre-deployment evaluations indicate the model is better than Sonnet 4.6 at refusing malicious requests, resisting prompt injection attacks, and reducing hallucinations and sycophancy. While it still lags behind Opus 4.8 on some alignment measures, Anthropic says Sonnet 5 is overall safer than its predecessor.

The model was not specifically trained for offensive cybersecurity tasks and launches with real-time cybersecurity safeguards enabled by default. These safeguards are designed to detect and block dangerous use. Developers can transition to Claude Sonnet 5 by switching from Claude Sonnet 4.6, with adaptive thinking enabled by default.