Anthropic just released Claude Opus 5 across the Claude apps, Claude Code, and the API, describing it as a thoughtful and proactive AI that nears the intelligence of its closest competitor at half the price. The new Opus 5 arrives with eval scores better than 4.8, as expected. What is surprising is that it even out-benches Fable 5 on many benchmarks, at a price that is hard to ignore.
Anthropic’s New Opus 5 Launches at Half the Fable Cost
The headline figure is simple: Opus 5 matches Fable 5 on key reasoning tasks while keeping the same $5 and $25 per million token pricing as Claude 4.8. That pricing removes the premium gap that previously forced teams to choose between capability and budget. For Australian developers and enterprises already watching AI costs, the move makes high-tier reasoning accessible without switching ecosystems.
Benchmark Dominance and Alignment
Anthropic says Opus 5 hits state-of-the-art 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 hit 30.2%, three times as high as the next best model, also ranking number one on the Artificial Analysis Intelligence Index leaderboard. On the International Math Olympiad 2026 problems, it scored a perfect 42 out of 42, well past the 29-point gold threshold.
The company also notes it is the most aligned AI in the family. It matches Mythos 5 at finding software bugs but stays well behind it at writing exploits. That gap matters for security teams evaluating model risk.
Why the Price Shift Matters for Developers
This is Anthropic filling Fable gaps with what it does best: offering intelligence from a tier above at a far lower price. So far, Sonnet has been the model that matched Opus. Now, Opus is matching Fable at the same per-token cost as 4.8, giving people who will not pay Fable rates a reliable and powerful option to work with. The shift should accelerate adoption in coding agents, research pipelines, and customer-service automation.