Anthropic released Claude Sonnet 5 — the most agentic model in the Sonnet line

Anthropic released Claude Sonnet 5 on June 30, 2026 — the most agentic model in the Sonnet lineup, which in quality approaches the flagship Claude Opus 4.8 but costs significantly less.

What the New Model Can Do

Claude Sonnet 5 can make plans, use tools like browsers and terminals, and work autonomously at a level that just a few months ago required larger and more expensive models. The Sonnet-tier models — Claude Sonnet 3.5, 3.6, and 3.7 — were the first to show impressive results in coding and tool use and essentially opened the era of agentic AI for developers. Recently, most of the gains in agentic capabilities have come from Opus-tier models, and Sonnet 5 should narrow this gap.

• Release date — June 30, 2026, available on all Claude pricing tiers
• Predecessor — Claude Sonnet 4.6, quality benchmark — Claude Opus 4.8
• API pricing at launch — $2 per million input tokens and $10 per million output tokens through August 31, 2026, then $3 and $15 respectively
• Default model in free and Pro tiers, available to Max, Team, and Enterprise users, as well as in Claude Code and Claude Platform under the name claude-sonnet-5
• Testing conducted on agent search benchmarks BrowseComp and computer interaction OSWorld-Verified

How Much Safer Is Sonnet 5 Than the Previous Model

According to Anthropic's assessment, Sonnet 5 shows a lower frequency of undesirable behavior than Sonnet 4.6 and is generally safer in agentic usage scenarios. However, its capability in cybersecurity tasks is noticeably lower than current Opus-tier models.

How Sonnet 5 Compares to Opus 4.8

In Anthropic's charts, Sonnet 5 shows consistent improvement over Sonnet 4.6 at different effort levels. Opus 4.8 remains the model of choice where maximum accuracy is needed, but Sonnet 5 gives developers significantly better results at a low price compared to what was previously available. Users can themselves adjust the effort level, choosing a balance between cost and accuracy — both within Sonnet 5 itself and when choosing between Sonnet 5 and Opus 4.8.

What It Means

Anthropic is systematically bringing the agentic capabilities of flagship models down to a more accessible and cheaper level: where six months ago an expensive Opus was required for autonomous agent work, now Sonnet is often sufficient — this lowers the barrier to entry for developers building products based on Claude.

Frequently Asked Questions

How

Much Does Access to Claude Sonnet 5 Cost Through the API?

Through August 31, 2026, the price is $2 per million input and $10 per million output tokens; after that date it will increase to $3 and $15 respectively.

How Does Sonnet 5 Differ From Opus 4.8?

Opus 4.8 remains more accurate on complex agentic tasks, but Sonnet 5 offers comparable quality at a significantly lower price; effort level can be adjusted to choose the balance between cost and accuracy.

Where Is Claude Sonnet 5 Available?

It became the default model in free and Pro tiers of Claude and is available to Max, Team, and Enterprise users, as well as in Claude Code and on Claude Platform via API under the name claude-sonnet-5.