Anthropic released Claude Opus 5.5 on September 22, saying the model matches Claude Fable 5.1 on most tasks while costing about 40% less to run than Opus 5 on typical workloads. List prices cited in coverage are $4 per million input tokens and $20 per million output tokens.
The company also reports faster generation — about 30% quicker output — and a jump on Terminal-Bench 4.0 to 66.4%, versus 52.3% for Opus 5. Those figures come from Anthropic’s launch materials as summarized by AI Weekly and BNN Bloomberg.
The timing is notable. Coverage notes this is Anthropic’s first major model drop after CEO Dario Amodei’s recent comments about pacing the frontier. The product still sits in the premium band, but the cost cut is aimed at customers who want Opus-class work without the previous bill.
Together with OpenAI’s cheaper GPT-6 Sol and Luna, the launch underscores a week defined less by raw capability leaps and more by a race to make frontier-adjacent models affordable enough for production agents.
Key takeaway. Opus 5.5 is Anthropic’s bid to keep premium quality while shrinking the price gap: near-Fable performance, faster output, and roughly 40% lower typical-workload cost than Opus 5.
Photo: Unsplash. Sources: Anthropic launch coverage via AI Weekly; BNN Bloomberg; Economic Times; Gizmodo (Sept. 22–23, 2026).
