Anthropic Releases Opus 5 With Near-Fable 5 Capabilities
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Claude Opus 5 on Thursday, its newest model, which the company said "comes close to the capabilities of Claude Fable 5 in many domains" and is significantly better at complex coding tasks.
- Opus 5 ships with stronger cybersecurity safeguards than its predecessor Opus 4.8, and Anthropic called it "the most aligned Opus model and is the least susceptible to being tricked into misuse" — a response to the government concerns that took Fable 5 and Mythos 5 offline for weeks.
- Anthropic is marketing Opus 5 to enterprise customers for knowledge work and biology as the ideal everyday model, while reserving Fable 5 for "the most ambitious types of work and long-horizon AI agent projects."
- Opus 5 pricing matches Opus 4.8 at $5 input / $25 output per million tokens, and Anthropic emphasized it is "half the price" of Fable 5 and slightly cheaper than OpenAI's GPT-5.6.
- Anthropic added a "Fast mode" option for Opus 5 launching in research preview at double the standard price, plus an option for automatic fallbacks to a lower-tier model when safeguards force the system to decline a request.
- Spokesperson Danielle Ghiglieri told The Verge that Anthropic "continues to work with our government partners to conduct their own independent testing of our models. This includes Opus 5."
Why it matters: Anthropic is deliberately segmenting its lineup: Opus 5 delivers near-frontier capability at mainstream pricing ($5/$25 per million tokens) while keeping regulator-scrutinized Fable 5 reserved for the most ambitious workloads — giving enterprise buyers a cheaper path that doesn't require the government's heightened cyber review, and undercutting OpenAI's GPT-5.6 on price.
