Anthropic Launches Opus 5.5 with Lower Pricing and Performance Rivaling Fable

Pricing and Performance Enhancements

Anthropic announced the release of its latest model, Opus 5.5, on Tuesday, positioning it as the new benchmark for coding and knowledge work tasks. The model sits at the top of the company’s three‑tier Claude lineup, above the mid‑range Sonnet and the fast, low‑cost Haiku. According to Anthropic, Opus 5.5 outperforms the larger Fable model on many benchmarks and succeeds on informal tasks where Fable failed.

In addition to performance gains, Opus 5.5 is markedly cheaper. The cost for output tokens is set at $20 per million tokens, down from $25 for the previous version. The price reduction extends to other metrics, and the model also requires less compute, resulting in faster execution.

The new version also refines its communication style. Opus 5.5 is less likely to use jargon and tends to place key information at the beginning of its responses.

Competitive Positioning and Safeguards

The launch follows the July 24 release of Opus 5 by just two months. Anthropic indicated that Sonnet 5.5 and Haiku 5.5 will arrive in the coming weeks with similar performance improvements.

Opus 5.5’s capabilities in biology and cybersecurity are comparable to the Mythos model, prompting the same safeguards that apply to the Fable model. These safeguards restrict the model’s use for discovering exploits in compiled programs or creating recognizable biological weapons, among other prohibited tasks.

Alignment and Safety Commitments

CEO Dario Amodei has embraced a strategy of pacing frontier progress. In a post earlier this month, he wrote,

I have become convinced that fully addressing the risks requires even more prudence, not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.

The safety training for Opus 5.5 mirrors that of its predecessors, incorporating alignment testing and pre‑release evaluation by external groups such as METR and Frontier Design. Anthropic noted that more advanced training and evaluation systems are already under development for future models, including enhanced security and monitoring tools.

Reflecting on the broader role of policy, the company stated,

As AI becomes more capable, public policy should play a larger role in making sure the systems people rely on are safe. That capacity takes time to build, and we’ve started to put the infrastructure in place to support it. We expect to share more details on these efforts soon.

Company Strategy and Future Roadmap

Opus 5.5 marks Anthropic’s first model release since CEO Dario Amodei pledged to deliberately slow AI capability progress to align with alignment efforts. The upcoming Sonnet 5.5 and Haiku 5.5 releases will continue this trajectory, delivering performance improvements while maintaining the company’s safety framework.