The Information Machine

Introducing Claude Sonnet 5

Anthropic News · 2026-06-30

Anthropic launches Claude Sonnet 5, a new agentic model with performance near Opus 4.8 at a lower price point, available across all plans starting June 30, 2026, at introductory pricing of $2 per million input tokens.

Open original ↗

Appears in

Extraction

Topics: model-releaseagentic-aiai-safetyllm-pricingcoding-agents

Claims

  • Claude Sonnet 5 delivers agentic performance close to Opus 4.8 at a substantially lower price than Opus-class models.
  • Sonnet 5 shows lower rates of hallucination, sycophancy, and undesirable behaviors than its predecessor Sonnet 4.6.
  • Sonnet 5 cannot develop full working cybersecurity exploits but shows slightly higher partial success rates than Sonnet 4.6 due to general intelligence gains, not specific training.
  • Sonnet 5 launches with introductory pricing of $2 per million input tokens and $10 per million output tokens through August 31, 2026, designed to be roughly cost-neutral relative to Sonnet 4.6 despite a new tokenizer.
  • Sonnet 5 is now the default model for Free and Pro Claude plans and is available in Claude Code and the Claude API.

Key quotes

Claude Sonnet 5 is built to be the most agentic Sonnet model yet. It can make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models.
I asked Claude Sonnet 5 to investigate a bug. Unprompted, it wrote a reproducing test, implemented the fix, then stashed it to confirm the bug came back without the change. All in a single pass.
A model that knows when to say no is just as important as one that knows how to build.