Daily AI CatchupClaude Opus 5 Launches Near Fable-5 Performance at Half the Price
Anthropic released Claude Opus 5 across the Claude apps, Claude Code, and the API — a model that nears Fable 5's intelligence at roughly half the price ($5/$25 per million tokens). It hit SOTA on agentic terminal coding, knowledge work, agentic search, and computer-use tasks, outperforming both Fable 5 and GPT-5.6 Sol. On ARC-AGI-3 it scored 30.2% — 3x the next-best model — and solved 42 of 42 International Math Olympiad 2026 problems, well past the gold threshold. Anthropic also disclosed that Opus 5 matches Mythos 5 at finding software bugs but lags well behind it at writing exploits.
Why it matters
💻 Developer · SOTA on agentic terminal coding and computer-use at half the price of comparable models is worth a direct swap-in test against whatever you're running today.
📦 Product · Near-flagship intelligence at half the usual cost changes the unit economics for any feature currently gated on frontier-model quality.
🎨 Design · Not directly relevant, though computer-use SOTA is worth watching if you're building any agent-driven UI automation.
📈 Business · A model that beats both Fable 5 and GPT-5.6 Sol on cost and benchmarks resets the competitive pricing conversation across the whole frontier-model market.
🤔 Just Curious · Anthropic released a new, smarter version of Claude called Opus 5 that performs almost as well as its best model but costs about half as much.
Sources: Anthropic: Introducing Claude Opus 5