Claude Sonnet 5: Features, Benchmarks, and Pricing (2026)

Claude Sonnet 5 is Anthropic's newest mid-tier model, launched June 30, 2026, and it's now the default Claude model for most users. It narrows the gap with the Opus class while staying faster and more affordable - built for developers and knowledge workers who need strong performance without frontier-model pricing. This guide is part of the AI Central Library, the resource 300,000+ senior professionals use to put AI to work.

Every model release claims to be smarter. Sonnet 5's pitch is more specific: it's Anthropic's most agentic Sonnet yet, running autonomously at a level that until recently required larger, more expensive models.

In practice that means near-Opus results on many tasks at a fraction of the cost - and a big step up from Sonnet 4.6 in reasoning, tool use, coding, and knowledge work.

Sonnet 5 isn't just better at answering questions. It's built for a fuller agentic workflow:

Early testers highlighted that it completes complex jobs where earlier Sonnet models would stop short, and checks its own output without being asked.

Sonnet 5 delivers major gains in sustained coding, tool coordination, debugging, and multi-step engineering tasks. It's built to handle real software workflows, not just generate code.

The headline numbers from Anthropic's launch benchmarks:

Note the last row: on knowledge work, Sonnet 5 actually edges past Opus 4.8 - a first for a Sonnet-class model.

Anthropic focused on improving how Claude thinks, not just what it knows. Sonnet 5 plans ahead, verifies its work, and follows through until completion. You can also adjust the effort level per task to find the right balance of cost and performance.

To get the most out of that reasoning, prompting still matters - our guide to prompting Claude beyond the basics pairs well with this release.

Claude Sonnet 5 is ideal for:

The launch partners tell the story. Zapier's team handed Sonnet 5 a two-part automation job - updating Salesforce account tiers, then sending a launch announcement to enterprise contacts - and it finished the whole thing end to end, where earlier models stalled halfway. As Zapier senior engineer Daniel Shepard put it: "For day-to-day automation, it's a no-brainer."

If you're setting Claude up for your team for the first time, start with our ChatGPT, Claude, and Copilot setup guide.

Anthropic also strengthened safety with this release. Sonnet 5 shows a lower overall rate of undesirable behaviors than Sonnet 4.6, with lower hallucination and sycophancy rates, cleaner refusals of malicious requests, and improved safeguards for agentic use.

Claude Sonnet 5 is available now across the Claude ecosystem: it's the default model on Free and Pro plans, available to Max, Team, and Enterprise users, and live in Claude Code and on the Claude Platform (API model string: claude-sonnet-5).

Introductory pricing (through August 31, 2026): $2 per 1M input tokens, $10 per 1M output tokens.

Standard pricing after that: $3 per 1M input tokens, $15 per 1M output tokens.

One caveat worth budgeting for: Sonnet 5 uses an updated tokenizer that maps the same text to roughly 1.0-1.35x more tokens than Sonnet 4.6. The introductory pricing was set to make the transition roughly cost-neutral, so re-test your real workloads before standard pricing kicks in.

AI models are no longer judged only by IQ - they're judged by execution. Can they finish the job? Claude Sonnet 5 is built for exactly that, and it sets the bar for what a mid-tier model should deliver in 2026.

Claude Sonnet 5 is Anthropic's mid-tier AI model, launched June 30, 2026. It's the most agentic Sonnet model to date - - able to plan, use tools, and complete multi-step tasks autonomously - and it's now the default model on Claude's Free and Pro plans.

Through August 31, 2026, Sonnet 5 has introductory API pricing of $2 per million input tokens and $10 per million output tokens. After that, standard pricing is $3 per million input and $15 per million output tokens.

Not across the board. Opus 4.8 still leads on agentic coding and computer use, but Sonnet 5 comes close on both and edges past Opus on knowledge work benchmarks - at a much lower price. For most everyday work, Sonnet 5 is the better value.

Sonnet 5 is substantially stronger in reasoning, tool use, coding, and knowledge work, and it's far more agentic - finishing complex multi-step tasks where Sonnet 4.6 would stall. It also shows lower rates of hallucination and other undesirable behaviors.