Anthropic launches Claude Opus 5 with near-Fable performance at half the cost

Anthropic has launched Claude Opus 5, its latest advanced AI model for complex coding, agentic workflows, scientific research, and knowledge work. It replaces Opus 4.8 at the same API pricing of $5 per million input tokens and $25 per million output tokens, while offering performance close to Claude Fable 5 at roughly half the cost for comparable workloads.

In benchmarks, Opus 5 scored 43.3% on Frontier-Bench v0.1, more than doubling Opus 4.8’s result, and achieved around three times the score of the next-best model on ARC-AGI 3. It also came close to Fable 5 on CursorBench 3.2 at half the cost and outperformed Fable 5 and GPT-5.6 Sol on several other benchmarks. Anthropic also reported stronger results at lower costs on Zapier AutomationBench and OSWorld 2.0.

Opus 5 is now the default model for Claude Max subscribers and the most capable option available to Claude Pro users. Developers can also change Claude’s tool permissions during active conversations without invalidating the prompt cache. An optional Fast mode delivers 2.5 times the speed at twice the standard price, and the model is available across Anthropic’s products and API under the identifier claude-opus-5.