Claude Opus 5 Lands, Approaching Fable 5 Performance at Half the Cost
Anthropic's Claude Opus 5 ships at unchanged Opus pricing while more than doubling Opus 4.8's score on Frontier-Bench and getting within 0.5% of Claude Fable 5 on coding at half the cost.
Anthropic has released Claude Opus 5, and the headline isn't just capability — it's price. Opus 5 ships at exactly the same price as Opus 4.8 ($5/$25 per million input/output tokens) while closing much of the gap to Anthropic's own top-tier Claude Fable 5, and in some benchmarks, beating everything else on the market outright.
What's under the hood
- 1-million-token context window, both default and maximum
- 128,000-token standard maximum output
- 96.0% on SWE-bench Verified
- Knowledge cutoff of May 2026 — the most current of any Claude model to date
- Now the default on Claude Max and the strongest model available on Claude Pro
How it actually performs
The benchmark story is the real headline:
| Benchmark | Result | |---|---| | Frontier-Bench v0.1 | 43.3% — more than double Opus 4.8's 18.7%, ahead of Fable 5's 33.7% | | CursorBench 3.2 | Within 0.5% of Fable 5's peak score, at roughly half the cost per task | | ARC-AGI 3 | Scores 3x the next-best model | | OSWorld 2.0 | Surpasses Fable 5's best result at roughly a third of the cost |
These benchmarks measure different things: Frontier-Bench and ARC-AGI test general reasoning, CursorBench tests real coding-agent tasks, and OSWorld tests computer-use — controlling a computer's UI directly. Opus 5 leading (or nearly matching) Fable 5 across all four, at a fraction of the cost on the ones where cost is measured, is what makes this a genuinely notable release rather than an incremental bump.
Why it matters
For developers and builders: frontier-class agentic coding and computer-use performance at unchanged Opus pricing is a real cost-efficiency jump — teams running Opus at scale get a substantially more capable model without a price increase.
For Anthropic's product lineup: Opus 5 narrows the gap to Fable 5 enough that the pricing tier, not just raw capability, becomes the main reason to choose one over the other — a genuinely different value calculus than most flagship-vs-flagship model comparisons.
For the competitive landscape: this lands the same month Google shipped a similarly priced Gemini 3.7 Flash update and Meta launched its own coding agent — frontier labs are now competing on price-per-capability, not just raw capability, across the board.
What to watch
- Independent benchmark verification now that Opus 5 is broadly available, beyond Anthropic's own reported numbers
- Whether Claude Fable 5's pricing shifts in response to Opus 5 narrowing the performance gap
- How Opus 5 performs in production agentic workloads, where benchmark gains don't always translate directly
Sources: MarkTechPost, Technology.org, EdTech Innovation Hub
NextGen AI Digest Editorial
Editorial Team
Reporting and analysis from the NextGen AI Digest newsroom — covering AI, agentic systems, SaaS, and the future of technology. Every piece is factual, sourced, and cited. Built and published by the team at Peaders.
Keep reading
Claude's Shared Chats and Artifacts Were Turning Up on Google
Some Claude conversations and Artifacts shared via public links have been indexed by Google, raising questions about the feature's privacy defaults.
Claude's Voice Mode Grows Up: Opus and Sonnet Replace Haiku-Only
Anthropic upgraded Claude's voice mode beyond Haiku for the first time, adding Opus and Sonnet model choice, cross-app connectors, and support for ten more languages.
The Fable 5 Reset: Why Anthropic's Most Powerful Model Went Dark and Came Back
Claude Fable 5 launched June 9, was forced offline by a U.S. export directive within three days, and returned with new cybersecurity classifiers. Inside the first government-ordered model recall.