← All news

Technology

Anthropic releases Claude Opus 5 at the same base token rates as Opus 4.8

Anthropic announced Claude Opus 5 on July 24, 2026, with stated base API prices unchanged from Opus 4.8 and access through Amazon Bedrock and Claude Platform on AWS.

reviewedUpdated Jul 26, 2026, 9:00 AM UTC
Original source

Anthropic

Read the original source

What happened

Anthropic listed Opus 5 at $5 per million input tokens and $25 per million output tokens. The Verge separately reported the same per-million-token rates and described them as unchanged from the predecessor.

AWS said customers can access Claude Opus 5 through Amazon Bedrock and Claude Platform on AWS. The reviewed sources establish the announced release and access channels, not universal regional availability, account eligibility, capacity, or actual user access.

Anthropic said Opus 5 improved performance at the same price and did not advance the frontier in risky dual-use capabilities, meaning capabilities that could support both beneficial and harmful uses.

Why it matters

The release pairs a new frontier-model capability claim with unchanged stated base token prices. That can affect the cost of AI tasks, but total task cost may also depend on output length, usage patterns, regional terms, discounts, rate limits, and Fast mode, which Anthropic says costs twice the base rate.

What remains unclear

Sources

  1. Introducing Claude Opus 5Primary source - Anthropic - company product announcement - Jul 24, 2026

    Used for: Release announcement, stated pricing, and vendor performance and safety claims.

    Open source

  2. Claude Opus 5 is now available on AWSAmazon Web Services - official cloud-provider availability announcement - Jul 24, 2026

    Used for: Availability through Amazon Bedrock and Claude Platform on AWS.

    Open source

  3. Anthropic releases Opus 5 with ‘close’ to Fable 5’s capabilitiesThe Verge - independent news report - Jul 24, 2026

    Used for: Independent reporting of the release, quoted base token prices, and attribution of capability and safety claims to Anthropic.

    Open source

  4. Artificial intelligenceNational Institute of Standards and Technology - government agency ai evaluation and risk-management overview - Date not listed

    Used for: General context on benchmarks, evaluations, and AI measurement science.

    Open source