Anthropic launches Claude Opus 5

Anthropic released Claude Opus 5, priced the same as its predecessor Opus 4.8 at $5 per million input tokens and $25 per million output tokens, available today via Claude Max, Claude Pro, and the API as claude-opus-5. It adds a low/medium/high effort toggle to balance cost against capability, plus an optional fast mode at 2x the base price that runs roughly 2.5x faster.

Why it matters

Opus 5 comes close to Claude Fable 5's frontier intelligence at roughly half the price, and by Anthropic's own benchmarks it's now the top performer on Frontier-Bench v0.1 (more than double Opus 4.8's score at a lower cost per task) and ARC-AGI 3 (roughly 3x the next-best model). It's not a universal win though: Anthropic's own reporting places it substantially behind Mythos 5 on cybersecurity exploitation tasks.

Who should care

Developers and teams doing agentic coding, business automation, or computer-use tasks who want frontier-level performance without frontier-level cost

What you can do

Switch to claude-opus-5 via the API, or it's already the default model on Claude Max and the strongest available on Claude Pro — try the new effort toggle to tune cost against capability for your workload.