Anthropic unveiled Opus 5 on July 24, 2026, as a new AI model positioned as a cheaper and less restrictive alternative to its existing Fable 5 model [1, 2, 3, 4, 5, 6]. Priced at $5 per million input tokens and $25 per million output tokens, Opus 5 costs about half as much as Fable 5, making it more accessible for users focused on cost-effective AI solutions [3, 4, 5, 6].
Opus 5 matches or slightly exceeds Fable 5’s performance in coding and knowledge work benchmarks such as Frontier-Bench, DeepSWE, GDPval-AA, and ARC-AGI 3. It also outperforms the earlier Opus 4.8 model and OpenAI’s GPT-5.6-Sol on most tasks [1, 2, 4, 6]. Anthropic stated Opus 5 is "much stronger at verifying its work and iterating carefully until it succeeds," citing benchmark tests [2].
The model includes reduced safety restrictions compared to Fable 5, with classifiers expected to intervene 85% less often. Anthropic said this allows more beneficial uses in areas like cybersecurity and biology [2, 5, 6]. "We expect the classifiers to intervene around 85 percent less often than they do for Fable 5," the company said [6].
Opus 5 can handle longer programming tasks with less human oversight due to improvements in verifying and iterating its output [2, 4]. It does not have a 30-day data retention policy that applies to Fable and Mythos models, appealing to users concerned about privacy [1, 2].
Anthropic deliberately avoided training Opus 5 on leading-edge cybersecurity attack and exploit generation tasks, which keeps it weaker than the Mythos 5 model in exploiting vulnerabilities, though it remains capable of finding them comparably well [1, 2, 3, 5, 6].
The company also launched a beta feature called Automatic Fallbacks that routes requests to a less powerful model if a safety classifier triggers. This reduces errors for API users relying on Opus 5 [2].
Dianne Penn, Anthropic’s head of product management for research, said, "Enterprises, in our feedback and with our customer base, are looking for value. If it's a cheaper model or a cheaper offering, but it's not accomplishing a similar level of quality, it's actually not useful" [3]. Anthropic positions Opus 5 as the model most users will choose for daily use due to its balance of cost, performance, and reduced restrictions [3, 6].
Earlier in 2026, Anthropic launched several models including Mythos 5, Fable 5, and Sonnet 5, and briefly disabled Fable 5 and Mythos 5 due to US export controls in June before restoring them after negotiations [2, 3, 5]. The company continues collaborating with government agencies on pre-deployment testing of Opus 5 and other models amid regulatory scrutiny [3, 5].
