Anthropic launched Claude Sonnet 5 on June 30, 2026, positioning it as the most capable and agentic Sonnet-class model to date. The company describes it as delivering performance close to its Opus 4.8 model while retaining the cost and speed characteristics expected of Sonnet-class offerings. Anthropic made the model available same-day across chat interfaces, coding tools, and its platform API.
Claude Sonnet 5 is built around agentic use cases, including coding, tool use, browser and terminal workflows, research, and professional-task automation. The model can form plans, operate autonomously across multi-step workflows, and handle long-horizon tasks at capability levels that previously required larger, more expensive models. Anthropic positions it as the default choice for everyday professional work, AI agents, and coding across its ecosystem.
Claude Sonnet 5 handles complex, multi-step agentic workflows — from coding to research — at a capability level once reserved for larger models.
The model ships with a 1 million token context window, which serves as both the default and maximum configuration. Maximum output length is set at 128,000 tokens, supporting extended generations for complex tasks and long-running agents. Reasoning is supported with configurable effort levels and adaptive thinking tied to higher-effort settings. The knowledge cutoff is January 2026. Claude Sonnet 5 uses the same tool interfaces as Claude Sonnet 4.6, with the exception that Priority Tier access is not available for the new model.
Anthropic introduced a promotional pricing structure through August 31, 2026, setting rates at $2 per million input tokens and $10 per million output tokens. After that period, standard pricing rises to $3 per million input tokens and $15 per million output tokens, aligning with Claude Sonnet 4.6 nominal rates. Anthropic notes up to 90% cost savings through prompt caching and up to 50% through batch processing.
One notable economic consideration is that Claude Sonnet 5 uses a new tokenizer that generates approximately 30% more tokens for the same input text compared to Claude Sonnet 4.6. Despite nominal price parity with Sonnet 4.6 at standard rates, the increased token count effectively raises the cost per unit of text processed.
Anthropic has also increased rate limits across Chat, Cowork, Claude Code, and Claude Platform to accommodate higher token volumes and elevated effort levels associated with the new model. The platform tiers have been simplified to three tiers on the native Claude Platform as part of this update.
Anthropic’s launch messaging frames Claude Sonnet 5 as the practical solution for teams and organizations deploying AI at scale. The emphasis falls on near-Opus intelligence delivered at a price and speed profile suitable for production agentic systems and routine professional workflows.
