Anthropic Launches Claude Sonnet 5 as Its New Default AI Model

Claude Sonnet 5 launch marks Anthropic’s latest update to its mid-range AI model lineup, with improved reasoning, coding, tool use, and agentic performance.

Anthropic has released Claude Sonnet 5 as the successor to Claude Sonnet 4.6. The new model skips the expected 4.x naming path and moves directly to version 5, signaling a larger update than a minor Sonnet refresh.

The company says Sonnet 5 closes much of the performance gap with Claude Opus 4.8 while keeping the Sonnet line positioned as a faster and more affordable option for everyday AI work.

Claude Sonnet 5 improves reasoning, coding, and tool use

Anthropic says Claude Sonnet 5 improves over Sonnet 4.6 across several important areas, including reasoning, agentic tasks, tool use, coding, and general knowledge work.

The model targets users who need strong performance for writing, research, software development, automation, and multi-step workflows. Anthropic also says Sonnet 5 should perform more reliably when using tools or working through longer agentic tasks.

That makes the update especially important for Claude Code users and developers building applications through the Claude API.

Claude Sonnet 5 still trails Opus 4.8 in advanced coding

Claude Sonnet 5 does not beat Opus 4.8 across the board.

In agentic coding, Sonnet 5 scores 63.2%, compared with 58.1% for Sonnet 4.6. That marks a clear improvement over the previous Sonnet model.

However, Opus 4.8 still leads in agentic coding with a score of 69.2%. For users handling the most complex coding tasks, Opus remains Anthropic’s stronger public model.

Sonnet 5 reportedly beats Opus 4.8 slightly on at least one knowledge-work benchmark, but Opus 4.8 still appears stronger for the most demanding reasoning and coding workloads.

Safer agentic workflows

Anthropic says Claude Sonnet 5 shows fewer undesirable behaviors than Sonnet 4.6.

The company describes the model as safer for agentic use cases, where AI systems complete multi-step tasks with tools or external context. Anthropic also says Sonnet 5 should hallucinate less and deliver more consistent results during agentic workflows.

Claude Sonnet 5 availability

Claude Sonnet 5 is rolling out across every Claude plan.

The model is now the default option for Free and Pro users. Max, Team, and Enterprise subscribers also get access.

Anthropic is also making Claude Sonnet 5 available through Claude Code and the Claude API, giving developers and enterprise customers direct access to the new model.

Claude Sonnet 5 pricing

Claude Sonnet 5 launches with lower temporary pricing than Sonnet 4.6.

The model costs $2 per million input tokens and $10 per million output tokens during the launch discount period. That pricing runs until August 31.

After the discount ends, Claude Sonnet 5 will cost $3 per million input tokens and $15 per million output tokens.

In other news, Anthropic’s ecosystem continues to expand as Claude models are now generally available in Microsoft Foundry on Azure. This move makes it easier for developers and enterprises to integrate Claude into their existing cloud workflows, bringing its capabilities for coding, reasoning, and automation directly into Microsoft’s platform.