Anthropic ships Claude Opus 5, tightens enterprise controls ahead of a pricing shift
A faster, cheaper Opus becomes the default on Claude Max, while Sonnet 5's promotional pricing ends August 31 and admin tooling gets deeper spend visibility.
Jordan Lee
AI Correspondent
Anthropic released Claude Opus 5, pitched as a faster and more cost-efficient model for coding, knowledge work, and scientific research. It's now the default model on Claude Max and the strongest option available on Claude Pro — a notable shift, since Opus-tier models have historically been reserved for higher-cost plans.
A pricing change is landing at the same time on the other end of the lineup: Claude Sonnet 5's promotional pricing of $2/$10 per million tokens ends August 31, 2026, with standard pricing of $3/$15 taking effect September 1. Teams that sized budgets around the promotional rate should re-check their projections before the switch.
Enterprise and compliance tooling
On the enterprise side, Anthropic added richer admin analytics, model-level entitlements, and spend alerts to Claude Enterprise, giving admins clearer visibility into usage, cost, and productivity trends without needing to build that reporting themselves. It also introduced inference hooks in beta, which let compliance teams enforce real-time data-loss-prevention rules across chat, Claude Code, and Cowork — inspecting prompts and tool calls before they ever reach the model.
Separately, subscription access to Claude Fable 5 was finalized on July 20: Max and Team Premium plans keep it included up to 50% of weekly usage limits, while Pro and Team Standard move to metered usage credits. The Developer Platform is also retiring its legacy Workbench and experimental prompt tools APIs, with access ending August 17, 2026.
More from AI News
OpenAI restructures around Sol/Terra/Luna tiers, ships ChatGPT Work
The new GPT-5.6 lineup comes with steep price cuts and an agentic system that runs multi-hour projects across a team's files and apps.
Google's AI Overviews now cover 43% of searches — and Search is getting its own agents
Coverage has nearly tripled in a year, AI Mode has passed a billion monthly users, and Google I/O 2026 introduced agents that work inside Search on your behalf.
Meta pivots from open Llama toward proprietary models with Muse Spark
The company that popularized open-weight frontier models is shifting course — its next flagship, Avocado, is delayed, and its first post-reorg release ships closed.