Anthropic releases Claude Opus 5

Background

Anthropic compared Opus 5 with Claude Fable 5, which it described as a model at the frontier of intelligence. ^[1]

What happened

On 24 July 2026 Anthropic released Claude Opus 5 at $5 per million input tokens and $25 per million output tokens. ^[1]

Anthropic said the model came close to the frontier intelligence of Claude Fable 5 at half the price. ^[1]

Key details

Anthropic reported state-of-the-art results on Frontier-Bench v0.1 and on GDPval-AA. It also reported that Opus 5 scored about three times the next-best model's score on ARC-AGI 3. ^[1]

Anthropic called Opus 5 its most aligned model to date, and its most capable generally available model for scientific research. ^[1]

Opus 5 became the new default model on Claude Max. ^[1]

Frontier intelligence refers to the most capable models available at a given time. A behavioural audit tests how a model acts in scenarios designed to reveal unwanted behaviour, and a misaligned-behaviour score summarises the results. Classifiers are systems that decide whether a request falls into a category that needs special handling.

Impact

Anthropic said its cyber classifiers were expected to intervene about 85% less often than those of Fable 5. ^[1]