Anthropic Opus 5.5 Launches With Lower Prices
Anthropic Opus 5.5 arrived on Tuesday, September 22, 2026, setting a new state-of-the-art in coding and knowledge work performance, according to the company. Opus is the most capable and expensive tier in Anthropic’s three-tier Claude lineup, with Sonnet in the middle and Haiku as the fastest and cheapest option.
Notably, Anthropic says Anthropic Opus 5.5 outpaces the larger Fable model in many benchmarks. It also succeeded in a number of informal tasks that Fable failed to complete.
A Significant Price Drop
The new model is significantly cheaper than its predecessor. Output tokens will be charged at $20 per million tokens, compared to $25 for the previous model. Other metrics have similar price drops.
The model is also faster to run, reflecting an overall drop in the compute required to serve it.
Changes to How Opus Communicates
Anthropic Opus 5.5 makes significant changes to how the model communicates. It is less likely to use jargon and more likely to put important information at the start of its messages.
A Quick Turnaround From Opus 5
The launch comes just two months after the release of Opus 5 on July 24. According to the announcement, Sonnet 5.5 and Haiku 5.5, the next tiers in the lineup, will be released “in the coming weeks,” with similar performance improvements.
Safety Safeguards and Capabilities
Anthropic says Anthropic Opus 5.5 is comparable to Mythos in its biology and cybersecurity capabilities. As a result, its release is subject to the same safeguards as the company’s Fable model.
Those safeguards limit how much the models can be used to discover exploits in compiled programs or develop recognizable biological weapons, among other tasks.
The First Release Under a New Pacing Approach
Anthropic Opus 5.5 is Anthropic’s first model release since CEO Dario Amodei embraced calls to pace the frontier. This means deliberately slowing down progress on AI capabilities to match the rate of progress on alignment.
“I have become convinced that fully addressing the risks requires even more prudence,” Amodei wrote in a post earlier this month. “Not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.”
Safety Training and External Evaluation
The safety training for Anthropic Opus 5.5 was broadly similar to its predecessors. It included alignment testing and pre-release evaluation by outside organizations like METR and Frontier Design.
Anthropic emphasized that more advanced training and evaluation systems were already being prepared for future models. These include improved security and monitoring systems.
“As AI becomes more capable, public policy should play a larger role in making sure the systems people rely on are safe,” the blog post reads. “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.”
What the Release Means for Users
For developers and businesses, Anthropic Opus 5.5 offers a rare combination: better performance and lower costs. The $5 drop in output token pricing, from $25 to $20 per million tokens, could meaningfully reduce expenses for high-volume workloads.
The speed improvements matter too. A model that requires less compute to serve is cheaper to run and faster to respond. That improves the experience for anyone building on top of it.
The communication changes may be the most noticeable difference for everyday users. Less jargon and more front-loaded answers make Anthropic Opus 5.5 easier to work with. This is especially true for people who use it as a writing or research assistant rather than a coding tool.
The Bigger Picture
Anthropic’s decision to release Anthropic Opus 5.5 just two months after Opus 5 shows the company is still shipping quickly. This is happening even as its CEO talks about pacing the frontier. The promised Sonnet 5.5 and Haiku 5.5 releases in the coming weeks suggest the whole Claude lineup is getting an upgrade cycle.
The comparison to Fable is also notable. Fable is a larger model, yet Anthropic claims Anthropic Opus 5.5 beats it on many benchmarks. If that holds up in independent testing, it reinforces a broader trend: smaller, more efficient models can compete with larger ones when trained and tuned well.
What to Watch Next
Independent benchmark results to verify Anthropic’s claims
The release timing and pricing for Sonnet 5.5 and Haiku 5.5
How the safety safeguards affect real-world use cases in biology and cybersecurity
Whether the pacing approach Amodei described slows future releases
For now, Anthropic Opus 5.5 is available as part of Anthropic’s Claude lineup. The company has not yet shared full details on the infrastructure efforts it mentioned in its blog post, but says more information is coming soon.
