Anthropic Launches Claude Sonnet 5 as Default for All Plans, With Agentic Tool Use and $2 Pricing

Sonnet 5 replaces 4.6 across every Claude plan today - with browser control, terminal use, and performance closer to Opus 4.8 at a fraction of the cost.

Saganote
Saganote ·
2 Min Read

Anthropic shipped Claude Sonnet 5 on June 30, replacing Sonnet 4.6 as the default model across all Claude plans - Free, Pro, Max, Team, and Enterprise. Agentic performance is the headline change: Sonnet 5 can plan multi-step tasks, use browsers and terminals, and run workflows that Anthropic says required larger models just a few months ago. Pricing starts at $2 per million input tokens.

Sonnet 5 Closes the Gap With Opus 4.8 on Agentic Tasks

Running Sonnet 5 costs a fraction of Opus 4.8. That gap matters to any team running agentic pipelines at scale - Claude Sonnet 5's performance is close enough to Opus 4.8 on reasoning, tool use, coding, and knowledge work that Anthropic now positions Sonnet as the right default rather than a fallback. For developers who spent months debating whether Sonnet 4.6 was good enough for production agentic work or whether they needed Opus, Sonnet 5 probably settles the question.

Anthropic also raised rate limits across Chat, Cowork, Claude Code, and the API alongside the launch. Rate limits up. Users can now select effort levels to control how much reasoning the model applies - lower effort for faster, cheaper responses; higher effort for complex multi-step tasks. Both changes address a friction point developers hit frequently with Sonnet 4.6.

Introductory Pricing at $2 Per Million Input Tokens Runs Through August 31

Standard pricing for Claude Sonnet 5 is $3 per million input tokens and $15 per million output tokens. Until August 31, Anthropic is offering it at $2 input and $10 output - a meaningful discount for teams building and testing agentic applications before committing to production infrastructure. Developers who have been running GPT-4o or similar mid-tier models for coding agents will find the pricing competitive without needing to model-shop.

Sonnet 5 is live now. Availability spans all Claude products and the API, with standard pricing kicking in September 1. OpenAI launched three GPT-5.6 models in June - Sol, Terra, and Luna - keeping mid-tier model competition between the two companies tight through the rest of 2026.


Share this
Saganote

About Author

Saganote

Saganote is an independent technology publication covering artificial intelligence, startups, cybersecurity, consumer technology, science, and innovation. Our editorial team reports on the companies, products, and ideas shaping the future.