Anthropic Releases Claude Opus 5, Claims Near-Fable 5 Intelligence at Half the Price

Anthropic released Claude Opus 5 on July 24, 2026, upgrading its flagship model line from version 4.8. The release arrives roughly two months after the previous Opus update, consistent with Anthropic's recent upgrade cadence.

What's new

Anthropic describes Opus 5 as a "thoughtful and proactive model that comes close to the frontier intelligence of Claude Fable 5 at half the price," according to a company statement posted on the Claude account on X. Fable 5 is Anthropic's most capable model — the publicly released, safeguarded version of a system internally codenamed Mythos, which Anthropic had delayed releasing over cybersecurity risk concerns.

Anthropic did not publish specific benchmark scores comparing Opus 5 to Fable 5 or competing models. The claim that Opus 5 approaches Fable 5-level intelligence is Anthropic's own characterization and has not been independently verified.

Opus 5 is now the default model on Claude Max and the strongest available model on Claude Pro, according to Anthropic. The company says the model is built for daily use and operates more efficiently than prior versions, though no specific latency or throughput figures were disclosed.

Pricing and availability

Claude Opus 5 is priced at $5 per million input tokens and $25 per million output tokens — identical to Opus 4.8 pricing. Anthropic did not raise prices despite the capability claims. The model is available today across all Claude platforms, including the API, Claude Pro, and Claude Max. Context window size was not disclosed in Anthropic's release materials.

Platform updates

Alongside the model release, Anthropic shipped two related platform features:

• Mid-conversation tool changes: Developers using the Claude Platform can now change which tools Claude has access to mid-conversation without invalidating the prompt cache, potentially reducing costs for long-running agent sessions.
• Automatic fallbacks on the API: Requests to Opus 5 or Fable 5 that get flagged by Anthropic's safety classifiers can now automatically route to another model instead of being blocked outright. With this setting enabled, API calls default to the best available model rather than failing on a safety flag.

What this means

Holding pricing flat at $5/$25 per million tokens while claiming near-flagship performance is a direct pitch to developers currently paying premium rates for Fable 5-tier output. If the intelligence gap is genuinely narrow, Opus 5 becomes the practical default for most production workloads, pushing Fable 5 into a smaller niche of cases that need maximum capability regardless of cost.

The automatic fallback feature is arguably the more consequential change for API users: false-positive safety blocks have long been a friction point for developers building agents, and silently downgrading to a fallback model rather than hard-failing a request changes the reliability calculus for production systems. Independent benchmarking against GPT and Gemini competitors will determine whether Anthropic's "near-Fable 5" claim holds up outside marketing copy.