Anthropic has introduced Claude Opus 5.5, the first model in its new Claude 5.5 family and a signal of the company's new strategic direction. The model achieves Claude Fable 5.1 performance while costing 40% less to operate than its predecessor Opus 5. This marks Anthropic's shift toward robustness and efficiency over raw capability growth.
Quick Facts
- Costs drop 40%: Input tokens cost $4, output tokens $20 per million (20% less than Opus 5); cache reads fall to $0.20 per million tokens (60% reduction)
- Speed increases 30%: Opus 5.5 generates output over 30% faster than its predecessor
- Safety first: The model scores highest in Anthropic's automated behavioral audit across all tested models; external evaluators from Frontier Design and METR tested it before release
- More natural communication: Testers report clearer, easier-to-follow writing—less of the typical "Claudish" style
Performance for Complex Tasks
Practical improvements are striking: one early tester completed a 680,000-line code migration in less than a day—work that would normally take an engineering team weeks. When optimizing web app load times, Opus 5.5 succeeded 39 of 40 times with genuine improvements, while Opus 5 made smaller gains that altered app behavior.
Creative tasks show similar progress: when multiple Claude models built a game from a single prompt, Opus 5.5 ranked highest for graphics quality and polish.
Safety with Verification Programs
Anthropic's safety approach sets the new model apart. Opus 5.5 achieves top scores on Anthropic's alignment suite—testing across thousands of simulated scenarios. The model is far less likely to take hard-to-reverse actions or exceed its boundaries, and resists prompt injection attacks better than Opus 5.
Because Opus 5.5 matches Claude Fable 5.1 in biology and cybersecurity capabilities, it deploys with similar safeguards. Organizations can now apply for Anthropic's Life Sciences Verification Program to use Opus 5.5 for biotech research. The Cyber Verification Program will expand in coming weeks, granting verified cybersecurity practitioners access.
Pricing Model and Limits
| Metric | Opus 5.5 | Opus 5 | Savings |
|---|---|---|---|
| Input tokens | $4/M | $5/M | 20% |
| Output tokens | $20/M | $25/M | 20% |
| Cache reads | $0.20/M | $0.50/M | 60% |
| Speed | +30% faster | Baseline | – |
Beyond price cuts, Anthropic raises five-hour usage limits on Pro, Max, Team, and Enterprise plans. Subscribers also receive a rate-limit reset they can use at any time.
What This Means for German Enterprises
The combination of cost efficiency, improved safety, and more natural communication makes Opus 5.5 attractive for German firms—especially those hesitant about high AI operating costs. Cybersecurity safeguards and verification programs could simplify regulatory compliance. More broadly, Anthropic's focus on robustness over speed signals an industry-wide trend: reliability is becoming a competitive advantage.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
