Anthropic Launches Claude Opus 5.5 with 40% Lower Costs and Enhanced Safeguards · Digg

Anthropic Launches Claude Opus 5.5 with 40% Lower Costs and Enhanced Safeguards

Anthropic released Claude Opus 5.5, reportedly matching prior frontier performance at significantly reduced inference costs (~40% lower overall, token prices cut ~20%), with faster output and stricter safeguards. The model emphasizes long-horizon coding and computer use capabilities.

1 Source, 32m ago, first seen 32m ago

TLDR

This release intensifies the AI model race amid prior "slowdown" narratives. Lower costs make frontier capabilities more accessible to developers and SaaS builders. The timing near OpenAI's cheaper GPT-6 variants fuels direct comparisons on value, safety, and competition. It marks the first major Anthropic release since discussions about pacing the frontier, signaling continued aggressive development despite prior caution rhetoric.

Featured Source

Anthropic has introduced Claude Opus 5.5, positioning the new model as a lower-cost option for demanding coding and agent work without presenting the release as a simple bargain-bin version of its prior flagship.

Anthropic says the model is available in Claude, Claude Code and the Claude Platform. Its listed API price is $4 per million input tokens and $20 per million output tokens, compared with $5 and $25 for Opus 5. The company also lists cache reads at $0.20 per million tokens, down from $0.50, a notable line item for long-running agents that repeatedly reuse context.

Lower cost is a company estimate, not a universal bill

Anthropic says Opus 5.5 costs 40% less than Opus 5 on typical workloads at default settings and generates output more than 30% faster. Those figures come from Anthropic’s own testing. Actual costs depend on the task, how much context is cached, the selected settings and the amount of tool use, so a developer’s bill will not necessarily track the headline percentage.

The launch also carries the company’s safety framing. Anthropic says it tested Opus 5.5 before release with Frontier Design and METR and is applying safeguards in areas including cybersecurity, biology and model distillation. It says certain cyber and biology work is routed through verification programs for vetted users, while general availability retains the broader restrictions.

For developers, the practical change is a pricing and deployment choice rather than a settled performance verdict. Anthropic’s benchmarks and early-tester accounts describe the model’s capabilities, but they are not a shared independent comparison with every competing system. Teams deciding whether to move workloads will still need to test quality, latency, tool behavior and total cost on the work they actually run.

Today's Rank

—

Not ranked yet

Anthropic Launches Claude Opus 5.5 with 40% Lower Costs and Enhanced Safeguards

1 Source, 32m ago, first seen 32m ago

TLDR

Featured Source

Lower cost is a company estimate, not a universal bill

1 Source

@ErSonusaini1⚡️ Claude Opus 5.5 is out — 40% cheaper $4/M in, $20/M out. • 66.4% Terminal-Bench 4.0 (vs 52.3% for Opus 5) • 81.8% OSWorld 2.0 computer use • 30% faster output • 85% less likely to attempt boundary circumvention First release since Amodei's "pace the frontier" essay. #AI #Claude #Anthropic

Today's Rank

—

Not ranked yet

Combined views

1 Source, first seen 32m ago

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.