Anthropic launched Claude Opus 5 on July 24, presenting it as their most affordable and efficient AI model yet. Priced at just half the cost of their previous Fable 5, Opus 5 charges $5 per million input tokens and $25 per million output tokens, making it a compelling choice for businesses aiming to manage AI expenses.
Unlike the premium Mythos 5, which remains Anthropic’s top-tier model and is reserved for select clients due to its advanced cybersecurity features, Opus 5 is positioned for everyday use. It delivers better performance than Fable 5 on coding and knowledge work benchmarks, striking a balance between capability and cost.
Competition and Market Dynamics
As enterprises tighten AI budgets, Anthropic’s head of product management, Dianne Penn, emphasized the need for models that offer both cost savings and high quality. This pressure comes amid fierce rivalry from tech giants like Microsoft, Amazon, and Google, alongside Chinese startups such as Moonshot AI. Moonshot AI recently introduced its Kimi K3, claiming performance on par with Fable 5 and OpenAI’s GPT-5.6 Sol.
The AI model race is intensifying alongside upcoming IPOs. Anthropic filed confidential paperwork in June and seems set to go public later this year. OpenAI, which rolled out its GPT-5.6 line earlier this month, might delay its IPO until 2027 amid a trade secrets lawsuit from Apple. Earlier in the year, Anthropic faced a temporary market withdrawal of Mythos 5 and Fable 5 due to cybersecurity concerns raised by the Trump administration, though access was restored after negotiations.
Anthropic continues collaboration with government agencies on pre-deployment testing, including for Opus 5, which remains behind Mythos 5 in offensive cybersecurity capabilities.
